Merge branch 'master' into devin/1765393705-fix-youtube-analytics-job-creation

.github/actions/connector-image-build-push/action.yml

@@ -139,8 +139,8 @@ runs:
           CONNECTOR_VERSION_TAG="${{ inputs.tag-override }}"
           echo "🏷 Using provided tag override: $CONNECTOR_VERSION_TAG"
         elif [[ "${{ inputs.release-type }}" == "pre-release" ]]; then
-          hash=$(git rev-parse --short=10 HEAD)
-          CONNECTOR_VERSION_TAG="${CONNECTOR_VERSION}-dev.${hash}"
+          hash=$(git rev-parse --short=7 HEAD)
+          CONNECTOR_VERSION_TAG="${CONNECTOR_VERSION}-preview.${hash}"
           echo "🏷 Using pre-release tag: $CONNECTOR_VERSION_TAG"
         else
           CONNECTOR_VERSION_TAG="$CONNECTOR_VERSION"

.github/pr-welcome-community.md

@@ -21,7 +21,7 @@ As needed or by request, Airbyte Maintainers can execute the following slash com
 - `/run-live-tests` - Runs live tests for the modified connector(s).
 - `/run-regression-tests` - Runs regression tests for the modified connector(s).
 - `/build-connector-images` - Builds and publishes a pre-release docker image for the modified connector(s).
-- `/publish-connectors-prerelease` - Publishes pre-release connector builds (tagged as `{version}-dev.{git-sha}`) for all modified connectors in the PR.
+- `/publish-connectors-prerelease` - Publishes pre-release connector builds (tagged as `{version}-preview.{git-sha}`) for all modified connectors in the PR.
 
 If you have any questions, feel free to ask in the PR comments or join our [Slack community](https://airbytehq.slack.com/).
 

.github/pr-welcome-internal.md

@@ -28,7 +28,11 @@ Airbyte Maintainers (that's you!) can execute the following slash commands on yo
 - `/run-live-tests` - Runs live tests for the modified connector(s).
 - `/run-regression-tests` - Runs regression tests for the modified connector(s).
 - `/build-connector-images` - Builds and publishes a pre-release docker image for the modified connector(s).
-- `/publish-connectors-prerelease` - Publishes pre-release connector builds (tagged as `{version}-dev.{git-sha}`) for all modified connectors in the PR.
+- `/publish-connectors-prerelease` - Publishes pre-release connector builds (tagged as `{version}-preview.{git-sha}`) for all modified connectors in the PR.
+- Connector release lifecycle (AI-powered):
+  - `/ai-prove-fix` - Runs prerelease readiness checks, including testing against customer connections.
+  - `/ai-canary-prerelease` - Rolls out prerelease to 5-10 connections for canary testing.
+  - `/ai-release-watch` - Monitors rollout post-release and tracks sync success rates.
 - JVM connectors:
   - `/update-connector-cdk-version connector=<CONNECTOR_NAME>` - Updates the specified connector to the latest CDK version.
     Example: `/update-connector-cdk-version connector=destination-bigquery`

.github/workflows/ai-canary-prerelease-command.yml

@@ -0,0 +1,72 @@
+name: AI Canary Prerelease Command
+
+on:
+  workflow_dispatch:
+    inputs:
+      pr:
+        description: "Pull request number (if triggered from a PR)"
+        type: number
+        required: false
+      comment-id:
+        description: "The comment-id of the slash command. Used to update the comment with the status."
+        required: false
+      repo:
+        description: "Repo (passed by slash command dispatcher)"
+        required: false
+        default: "airbytehq/airbyte"
+      gitref:
+        description: "Git ref (passed by slash command dispatcher)"
+        required: false
+
+run-name: "AI Canary Prerelease for PR #${{ github.event.inputs.pr }}"
+
+permissions:
+  contents: read
+  issues: write
+  pull-requests: read
+
+jobs:
+  ai-canary-prerelease:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Get job variables
+        id: job-vars
+        run: |
+          echo "run-url=https://github.com/$GITHUB_REPOSITORY/actions/runs/$GITHUB_RUN_ID" >> $GITHUB_OUTPUT
+
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Authenticate as GitHub App
+        uses: actions/create-github-app-token@v2
+        id: get-app-token
+        with:
+          owner: "airbytehq"
+          repositories: "airbyte,oncall"
+          app-id: ${{ secrets.OCTAVIA_BOT_APP_ID }}
+          private-key: ${{ secrets.OCTAVIA_BOT_PRIVATE_KEY }}
+
+      - name: Post start comment
+        if: inputs.comment-id != ''
+        uses: peter-evans/create-or-update-comment@v4
+        with:
+          token: ${{ steps.get-app-token.outputs.token }}
+          comment-id: ${{ inputs.comment-id }}
+          issue-number: ${{ inputs.pr }}
+          body: |
+            > **AI Canary Prerelease Started**
+            >
+            > Rolling out to 5-10 connections, watching results, and reporting findings.
+            > [View workflow run](${{ steps.job-vars.outputs.run-url }})
+
+      - name: Run AI Canary Prerelease
+        uses: aaronsteers/devin-action@main
+        with:
+          comment-id: ${{ inputs.comment-id }}
+          issue-number: ${{ inputs.pr }}
+          playbook-macro: "!canary_prerelease"
+          devin-token: ${{ secrets.DEVIN_AI_API_KEY }}
+          github-token: ${{ steps.get-app-token.outputs.token }}
+          start-message: "🐤 **AI Canary Prerelease session starting...** Rolling out to 5-10 connections, watching results, and reporting findings. [View playbook](https://github.com/airbytehq/oncall/blob/main/prompts/playbooks/canary_prerelease.md)"
+          tags: |
+            ai-oncall

.github/workflows/ai-prove-fix-command.yml

@@ -0,0 +1,72 @@
+name: AI Prove Fix Command
+
+on:
+  workflow_dispatch:
+    inputs:
+      pr:
+        description: "Pull request number (if triggered from a PR)"
+        type: number
+        required: false
+      comment-id:
+        description: "The comment-id of the slash command. Used to update the comment with the status."
+        required: false
+      repo:
+        description: "Repo (passed by slash command dispatcher)"
+        required: false
+        default: "airbytehq/airbyte"
+      gitref:
+        description: "Git ref (passed by slash command dispatcher)"
+        required: false
+
+run-name: "AI Prove Fix for PR #${{ github.event.inputs.pr }}"
+
+permissions:
+  contents: read
+  issues: write
+  pull-requests: read
+
+jobs:
+  ai-prove-fix:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Get job variables
+        id: job-vars
+        run: |
+          echo "run-url=https://github.com/$GITHUB_REPOSITORY/actions/runs/$GITHUB_RUN_ID" >> $GITHUB_OUTPUT
+
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Authenticate as GitHub App
+        uses: actions/create-github-app-token@v2
+        id: get-app-token
+        with:
+          owner: "airbytehq"
+          repositories: "airbyte,oncall"
+          app-id: ${{ secrets.OCTAVIA_BOT_APP_ID }}
+          private-key: ${{ secrets.OCTAVIA_BOT_PRIVATE_KEY }}
+
+      - name: Post start comment
+        if: inputs.comment-id != ''
+        uses: peter-evans/create-or-update-comment@v4
+        with:
+          token: ${{ steps.get-app-token.outputs.token }}
+          comment-id: ${{ inputs.comment-id }}
+          issue-number: ${{ inputs.pr }}
+          body: |
+            > **AI Prove Fix Started**
+            >
+            > Running readiness checks and testing against customer connections.
+            > [View workflow run](${{ steps.job-vars.outputs.run-url }})
+
+      - name: Run AI Prove Fix
+        uses: aaronsteers/devin-action@main
+        with:
+          comment-id: ${{ inputs.comment-id }}
+          issue-number: ${{ inputs.pr }}
+          playbook-macro: "!prove_fix"
+          devin-token: ${{ secrets.DEVIN_AI_API_KEY }}
+          github-token: ${{ steps.get-app-token.outputs.token }}
+          start-message: "🔍 **AI Prove Fix session starting...** Running readiness checks and testing against customer connections. [View playbook](https://github.com/airbytehq/oncall/blob/main/prompts/playbooks/prove_fix.md)"
+          tags: |
+            ai-oncall

.github/workflows/ai-release-watch-command.yml

@@ -0,0 +1,72 @@
+name: AI Release Watch Command
+
+on:
+  workflow_dispatch:
+    inputs:
+      pr:
+        description: "Pull request number (if triggered from a PR)"
+        type: number
+        required: false
+      comment-id:
+        description: "The comment-id of the slash command. Used to update the comment with the status."
+        required: false
+      repo:
+        description: "Repo (passed by slash command dispatcher)"
+        required: false
+        default: "airbytehq/airbyte"
+      gitref:
+        description: "Git ref (passed by slash command dispatcher)"
+        required: false
+
+run-name: "AI Release Watch for PR #${{ github.event.inputs.pr }}"
+
+permissions:
+  contents: read
+  issues: write
+  pull-requests: read
+
+jobs:
+  ai-release-watch:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Get job variables
+        id: job-vars
+        run: |
+          echo "run-url=https://github.com/$GITHUB_REPOSITORY/actions/runs/$GITHUB_RUN_ID" >> $GITHUB_OUTPUT
+
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Authenticate as GitHub App
+        uses: actions/create-github-app-token@v2
+        id: get-app-token
+        with:
+          owner: "airbytehq"
+          repositories: "airbyte,oncall"
+          app-id: ${{ secrets.OCTAVIA_BOT_APP_ID }}
+          private-key: ${{ secrets.OCTAVIA_BOT_PRIVATE_KEY }}
+
+      - name: Post start comment
+        if: inputs.comment-id != ''
+        uses: peter-evans/create-or-update-comment@v4
+        with:
+          token: ${{ steps.get-app-token.outputs.token }}
+          comment-id: ${{ inputs.comment-id }}
+          issue-number: ${{ inputs.pr }}
+          body: |
+            > **AI Release Watch Started**
+            >
+            > Monitoring rollout and tracking sync success rates.
+            > [View workflow run](${{ steps.job-vars.outputs.run-url }})
+
+      - name: Run AI Release Watch
+        uses: aaronsteers/devin-action@main
+        with:
+          comment-id: ${{ inputs.comment-id }}
+          issue-number: ${{ inputs.pr }}
+          playbook-macro: "!release_watch"
+          devin-token: ${{ secrets.DEVIN_AI_API_KEY }}
+          github-token: ${{ steps.get-app-token.outputs.token }}
+          start-message: "👁️ **AI Release Watch session starting...** Monitoring rollout and tracking sync success rates. [View playbook](https://github.com/airbytehq/oncall/blob/main/prompts/playbooks/release_watch.md)"
+          tags: |
+            ai-oncall

.github/workflows/autodoc.yml

@@ -104,7 +104,7 @@ jobs:
         if: steps.check-support-level.outputs.metadata_file == 'true' && steps.check-support-level.outputs.community_support == 'true'
         env:
           PROMPT_TEXT: "The commit to review is ${{ github.sha }}. This commit was pushed to master and may contain connector changes that need documentation updates."
-        uses: aaronsteers/devin-action@0d74d6d9ff1b16ada5966dc31af53a9d155759f4 # Pinned to specific commit for security
+        uses: aaronsteers/devin-action@98d15ae93d1848914f5ab8e9ce45341182958d27 # v0.1.7 - Pinned to specific commit for security
         with:
           devin-token: ${{ secrets.DEVIN_AI_API_KEY }}
           github-token: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/label-community-prs.yml

@@ -0,0 +1,28 @@
+name: Label Community PRs
+
+# This workflow automatically adds the "community" label to PRs from forks.
+# This enables automatic tracking on the Community PRs project board.
+
+on:
+  pull_request_target:
+    types:
+      - opened
+      - reopened
+
+jobs:
+  label-community-pr:
+    name: Add "Community" Label to PR
+    # Only run for PRs from forks
+    if: github.event.pull_request.head.repo.fork == true
+    runs-on: ubuntu-24.04
+    permissions:
+      issues: write
+      pull-requests: write
+    steps:
+      - name: Add community label
+        # This action uses GitHub's addLabels API, which is idempotent.
+        # If the label already exists, the API call succeeds without error.
+        uses: actions-ecosystem/action-add-labels@bd52874380e3909a1ac983768df6976535ece7f8 # v1.1.3
+        with:
+          github_token: ${{ secrets.GITHUB_TOKEN }}
+          labels: community

.github/workflows/publish-connectors-prerelease-command.yml

@@ -3,7 +3,7 @@ name: Publish Connectors Pre-release
 # It can be triggered via the /publish-connectors-prerelease slash command from PR comments,
 # or via the MCP tool `publish_connector_to_airbyte_registry`.
 #
-# Pre-release versions are tagged with the format: {version}-dev.{10-char-git-sha}
+# Pre-release versions are tagged with the format: {version}-preview.{7-char-git-sha}
 # These versions are NOT eligible for semver auto-advancement but ARE available
 # for version pinning via the scoped_configuration API.
 #
@@ -66,7 +66,7 @@ jobs:
       - name: Get short SHA
         id: get-sha
         run: |
-          SHORT_SHA=$(git rev-parse --short=10 HEAD)
+          SHORT_SHA=$(git rev-parse --short=7 HEAD)
           echo "short-sha=$SHORT_SHA" >> $GITHUB_OUTPUT
 
       - name: Get job variables
@@ -135,7 +135,7 @@ jobs:
             > Publishing pre-release build for connector `${{ steps.resolve-connector.outputs.connector-name }}`.
             > Branch: `${{ inputs.gitref }}`
             >
-            > Pre-release versions will be tagged as `{version}-dev.${{ steps.get-sha.outputs.short-sha }}`
+            > Pre-release versions will be tagged as `{version}-preview.${{ steps.get-sha.outputs.short-sha }}`
             > and are available for version pinning via the scoped_configuration API.
             >
             > [View workflow run](${{ steps.job-vars.outputs.run-url }})
@@ -147,6 +147,7 @@ jobs:
     with:
       connectors: ${{ format('--name={0}', needs.init.outputs.connector-name) }}
       release-type: pre-release
+      gitref: ${{ inputs.gitref }}
     secrets: inherit
 
   post-completion:
@@ -176,13 +177,12 @@ jobs:
         id: message-vars
         run: |
           CONNECTOR_NAME="${{ needs.init.outputs.connector-name }}"
-          SHORT_SHA="${{ needs.init.outputs.short-sha }}"
-          VERSION="${{ needs.init.outputs.connector-version }}"
+          # Use the actual docker-image-tag from the publish workflow output
+          DOCKER_TAG="${{ needs.publish.outputs.docker-image-tag }}"
 
-          if [[ -n "$VERSION" ]]; then
-            DOCKER_TAG="${VERSION}-dev.${SHORT_SHA}"
-          else
-            DOCKER_TAG="{version}-dev.${SHORT_SHA}"
+          if [[ -z "$DOCKER_TAG" ]]; then
+            echo "::error::docker-image-tag output is missing from publish workflow. This is unexpected."
+            exit 1
           fi
 
           echo "connector_name=$CONNECTOR_NAME" >> $GITHUB_OUTPUT

.github/workflows/publish_connectors.yml

@@ -21,6 +21,14 @@ on:
         required: false
         default: false
         type: boolean
+      gitref:
+        description: "Git ref (branch or SHA) to build connectors from. Used by pre-release workflow to build from PR branches."
+        required: false
+        type: string
+    outputs:
+      docker-image-tag:
+        description: "Docker image tag used when publishing. For single-connector callers only; multi-connector callers should not rely on this output."
+        value: ${{ jobs.publish_connector_registry_entries.outputs.docker-image-tag }}
   workflow_dispatch:
     inputs:
       connectors:
@@ -48,6 +56,7 @@ jobs:
         # v4
         uses: actions/checkout@08eba0b27e820071cde6df949e0beb9ba4906955
         with:
+          ref: ${{ inputs.gitref || '' }}
           fetch-depth: 2 # Required so we can conduct a diff from the previous commit to understand what connectors have changed.
           submodules: true # Required for the enterprise repo since it uses a submodule that needs to exist for this workflow to run successfully.
       - name: List connectors to publish [manual]
@@ -105,6 +114,7 @@ jobs:
         # v4
         uses: actions/checkout@08eba0b27e820071cde6df949e0beb9ba4906955
         with:
+          ref: ${{ inputs.gitref || '' }}
           fetch-depth: 2 # Required so we can conduct a diff from the previous commit to understand what connectors have changed.
           submodules: true # Required for the enterprise repo since it uses a submodule that needs to exist for this workflow to run successfully.
 
@@ -250,11 +260,14 @@ jobs:
       max-parallel: 5
       # Allow all jobs to run, even if one fails
       fail-fast: false
+    outputs:
+      docker-image-tag: ${{ steps.connector-metadata.outputs.docker-image-tag }}
     steps:
       - name: Checkout Airbyte
         # v4
         uses: actions/checkout@08eba0b27e820071cde6df949e0beb9ba4906955
         with:
+          ref: ${{ inputs.gitref || '' }}
           fetch-depth: 2 # Required so we can conduct a diff from the previous commit to understand what connectors have changed.
           submodules: true # Required for the enterprise repo since it uses a submodule that needs to exist for this workflow to run successfully.
 
@@ -292,8 +305,8 @@ jobs:
           echo "connector-version=$(poe -qq get-version)" | tee -a $GITHUB_OUTPUT
           CONNECTOR_VERSION=$(poe -qq get-version)
           if [[ "${{ inputs.release-type }}" == "pre-release" ]]; then
-            hash=$(git rev-parse --short=10 HEAD)
-            echo "docker-image-tag=${CONNECTOR_VERSION}-dev.${hash}" | tee -a $GITHUB_OUTPUT
+            hash=$(git rev-parse --short=7 HEAD)
+            echo "docker-image-tag=${CONNECTOR_VERSION}-preview.${hash}" | tee -a $GITHUB_OUTPUT
             echo "release-type-flag=--pre-release" | tee -a $GITHUB_OUTPUT
           else
             echo "docker-image-tag=${CONNECTOR_VERSION}" | tee -a $GITHUB_OUTPUT
@@ -349,6 +362,7 @@ jobs:
         # v4
         uses: actions/checkout@08eba0b27e820071cde6df949e0beb9ba4906955
         with:
+          ref: ${{ inputs.gitref || '' }}
           submodules: true # Required for the enterprise repo since it uses a submodule that needs to exist for this workflow to run successfully.
       - name: Match GitHub User to Slack User
         id: match-github-to-slack-user
@@ -381,6 +395,7 @@ jobs:
         # v4
         uses: actions/checkout@08eba0b27e820071cde6df949e0beb9ba4906955
         with:
+          ref: ${{ inputs.gitref || '' }}
           submodules: true # Required for the enterprise repo since it uses a submodule that needs to exist for this workflow to run successfully.
       - name: Notify PagerDuty
         id: pager-duty

.github/workflows/slash-commands.yml

@@ -35,6 +35,9 @@ jobs:
           issue-type: both
 
           commands: |
+            ai-canary-prerelease
+            ai-prove-fix
+            ai-release-watch
             approve-regression-tests
             bump-bulk-cdk-version
             bump-progressive-rollout-version

.github/workflows/sync-ai-connector-docs.yml

@@ -0,0 +1,70 @@
+name: Sync Agent Connector Docs
+
+on:
+  schedule:
+    - cron: "0 */2 * * *" # Every 2 hours
+  workflow_dispatch: # Manual trigger
+
+jobs:
+  sync-docs:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout airbyte repo
+        uses: actions/checkout@08eba0b27e820071cde6df949e0beb9ba4906955 # v4.3.0
+
+      - name: Checkout airbyte-agent-connectors
+        uses: actions/checkout@08eba0b27e820071cde6df949e0beb9ba4906955 # v4.3.0
+        with:
+          repository: airbytehq/airbyte-agent-connectors
+          path: agent-connectors-source
+
+      - name: Sync connector docs
+        run: |
+          DEST_DIR="docs/ai-agents/connectors"
+          mkdir -p "$DEST_DIR"
+
+          for connector_dir in agent-connectors-source/connectors/*/; do
+            connector=$(basename "$connector_dir")
+
+            # Only delete/recreate the specific connector subdirectory
+            # This leaves any files directly in $DEST_DIR untouched
+            rm -rf "$DEST_DIR/$connector"
+            mkdir -p "$DEST_DIR/$connector"
+
+            # Copy all markdown files for this connector
+            for md_file in "$connector_dir"/*.md; do
+              if [ -f "$md_file" ]; then
+                cp "$md_file" "$DEST_DIR/$connector/"
+              fi
+            done
+          done
+
+          echo "Synced $(ls -d $DEST_DIR/*/ 2>/dev/null | wc -l) connectors"
+
+      - name: Cleanup temporary checkout
+        run: rm -rf agent-connectors-source
+
+      - name: Authenticate as GitHub App
+        uses: actions/create-github-app-token@v2
+        id: get-app-token
+        with:
+          owner: "airbytehq"
+          repositories: "airbyte"
+          app-id: ${{ secrets.OCTAVIA_BOT_APP_ID }}
+          private-key: ${{ secrets.OCTAVIA_BOT_PRIVATE_KEY }}
+
+      - name: Create PR if changes
+        uses: peter-evans/create-pull-request@0979079bc20c05bbbb590a56c21c4e2b1d1f1bbe # v6
+        with:
+          token: ${{ steps.get-app-token.outputs.token }}
+          commit-message: "docs: sync agent connector docs from airbyte-agent-connectors repo"
+          branch: auto-sync-ai-connector-docs
+          delete-branch: true
+          title: "docs: sync agent connector docs from airbyte-agent-connectors repo"
+          body: |
+            Automated sync of agent connector docs from airbyte-agent-connectors.
+
+            This PR was automatically created by the sync-agent-connector-docs workflow.
+          labels: |
+            documentation
+            auto-merge

.markdownlintignore

@@ -0,0 +1,3 @@
+# Ignore auto-generated connector documentation files synced from airbyte-agent-connectors repo
+# These files are generated and have formatting that doesn't conform to markdownlint rules
+docs/ai-agents/connectors/**

airbyte-cdk/bulk/changelog.md

@@ -1,3 +1,21 @@
+## Version 0.1.91
+
+load cdk: upsert records test uses proper target schema
+
+## Version 0.1.90
+
+load cdk: components tests: data coercion tests cover all data types
+
+## Version 0.1.89
+
+load cdk: components tests: data coercion tests for int+number
+
+## Version 0.1.88
+
+**Load CDK**
+
+* Add CDC_CURSOR_COLUMN_NAME constant.
+
 ## Version 0.1.87
 
 **Load CDK**

airbyte-cdk/bulk/core/load/src/main/kotlin/io/airbyte/cdk/load/table/ColumnsConstants.kt

@@ -4,4 +4,13 @@
 
 package io.airbyte.cdk.load.table
 
+/**
+ * CDC meta column names.
+ *
+ * Note: These CDC column names are brittle as they are separate yet coupled to the logic sources
+ * use to generate these column names. See
+ * [io.airbyte.integrations.source.mssql.MsSqlSourceOperations.MsSqlServerCdcMetaFields] for an
+ * example.
+ */
 const val CDC_DELETED_AT_COLUMN = "_ab_cdc_deleted_at"
+const val CDC_CURSOR_COLUMN = "_ab_cdc_cursor"

airbyte-cdk/bulk/core/load/src/testFixtures/kotlin/io/airbyte/cdk/load/component/DataCoercionFixtures.kt

@@ -0,0 +1,859 @@
+/*
+ * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
+ */
+
+package io.airbyte.cdk.load.component
+
+import io.airbyte.cdk.load.data.AirbyteValue
+import io.airbyte.cdk.load.data.ArrayValue
+import io.airbyte.cdk.load.data.DateValue
+import io.airbyte.cdk.load.data.IntegerValue
+import io.airbyte.cdk.load.data.NullValue
+import io.airbyte.cdk.load.data.NumberValue
+import io.airbyte.cdk.load.data.ObjectValue
+import io.airbyte.cdk.load.data.StringValue
+import io.airbyte.cdk.load.data.TimeWithTimezoneValue
+import io.airbyte.cdk.load.data.TimeWithoutTimezoneValue
+import io.airbyte.cdk.load.data.TimestampWithTimezoneValue
+import io.airbyte.cdk.load.data.TimestampWithoutTimezoneValue
+import io.airbyte.cdk.load.dataflow.transform.ValueCoercer
+import io.airbyte.cdk.load.util.serializeToString
+import io.airbyte.protocol.models.v0.AirbyteRecordMessageMetaChange.Reason
+import java.math.BigDecimal
+import java.math.BigInteger
+import java.time.LocalDate
+import java.time.LocalDateTime
+import java.time.OffsetDateTime
+import java.time.format.DateTimeFormatter
+import java.time.format.DateTimeFormatterBuilder
+import java.time.format.SignStyle
+import java.time.temporal.ChronoField
+import org.junit.jupiter.params.provider.Arguments
+
+/*
+ * This file defines "interesting values" for all data types, along with expected behavior for those values.
+ * You're free to define your own values/behavior depending on the destination, but it's recommended
+ * that you try to match behavior to an existing fixture.
+ *
+ * Classes also include some convenience functions for JUnit. For example, you could annotate your
+ * method with:
+ * ```kotlin
+ * @ParameterizedTest
+ * @MethodSource("io.airbyte.cdk.load.component.DataCoercionIntegerFixtures#int64")
+ * ```
+ *
+ * By convention, all fixtures are declared as:
+ * 1. One or more `val <name>: List<Pair<AirbyteValue, Any?>>` (each pair representing the input value,
+ *    and the expected output value)
+ * 2. One or more `fun <name>(): List<Arguments> = <name>.toArgs()`, which can be provided to JUnit's MethodSource
+ *
+ * If you need to mutate fixtures in some way, you should reference the `val`, and use the `toArgs()`
+ * extension function to convert it to JUnit's Arguments class. See [DataCoercionIntegerFixtures.int64AsBigInteger]
+ * for an example.
+ */
+
+object DataCoercionIntegerFixtures {
+    // "9".repeat(38)
+    val numeric38_0Max = bigint("99999999999999999999999999999999999999")
+    val numeric38_0Min = bigint("-99999999999999999999999999999999999999")
+
+    const val ZERO = "0"
+    const val ONE = "1"
+    const val NEGATIVE_ONE = "-1"
+    const val FORTY_TWO = "42"
+    const val NEGATIVE_FORTY_TWO = "-42"
+    const val INT32_MAX = "int32 max"
+    const val INT32_MIN = "int32 min"
+    const val INT32_MAX_PLUS_ONE = "int32_max + 1"
+    const val INT32_MIN_MINUS_ONE = "int32_min - 1"
+    const val INT64_MAX = "int64 max"
+    const val INT64_MIN = "int64 min"
+    const val INT64_MAX_PLUS_ONE = "int64_max + 1"
+    const val INT64_MIN_MINUS_1 = "int64_min - 1"
+    const val NUMERIC_38_0_MAX = "numeric(38,0) max"
+    const val NUMERIC_38_0_MIN = "numeric(38,0) min"
+    const val NUMERIC_38_0_MAX_PLUS_ONE = "numeric(38,0)_max + 1"
+    const val NUMERIC_38_0_MIN_MINUS_ONE = "numeric(38,0)_min - 1"
+
+    /**
+     * Many destinations use int64 to represent integers. In this case, we null out any value beyond
+     * Long.MIN/MAX_VALUE.
+     */
+    val int64 =
+        listOf(
+            case(NULL, NullValue, null),
+            case(ZERO, IntegerValue(0), 0L),
+            case(ONE, IntegerValue(1), 1L),
+            case(NEGATIVE_ONE, IntegerValue(-1), -1L),
+            case(FORTY_TWO, IntegerValue(42), 42L),
+            case(NEGATIVE_FORTY_TWO, IntegerValue(-42), -42L),
+            // int32 bounds, and slightly out of bounds
+            case(INT32_MAX, IntegerValue(Integer.MAX_VALUE.toLong()), Integer.MAX_VALUE.toLong()),
+            case(INT32_MIN, IntegerValue(Integer.MIN_VALUE.toLong()), Integer.MIN_VALUE.toLong()),
+            case(
+                INT32_MAX_PLUS_ONE,
+                IntegerValue(Integer.MAX_VALUE.toLong() + 1),
+                Integer.MAX_VALUE.toLong() + 1
+            ),
+            case(
+                INT32_MIN_MINUS_ONE,
+                IntegerValue(Integer.MIN_VALUE.toLong() - 1),
+                Integer.MIN_VALUE.toLong() - 1
+            ),
+            // int64 bounds, and slightly out of bounds
+            case(INT64_MAX, IntegerValue(Long.MAX_VALUE), Long.MAX_VALUE),
+            case(INT64_MIN, IntegerValue(Long.MIN_VALUE), Long.MIN_VALUE),
+            // values out of int64 bounds are nulled
+            case(
+                INT64_MAX_PLUS_ONE,
+                IntegerValue(bigint(Long.MAX_VALUE) + BigInteger.ONE),
+                null,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION
+            ),
+            case(
+                INT64_MIN_MINUS_1,
+                IntegerValue(bigint(Long.MIN_VALUE) - BigInteger.ONE),
+                null,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION
+            ),
+            // NUMERIC(38, 9) bounds, and slightly out of bounds
+            // (these are all out of bounds for an int64 value, so they all get nulled)
+            case(
+                NUMERIC_38_0_MAX,
+                IntegerValue(numeric38_0Max),
+                null,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION
+            ),
+            case(
+                NUMERIC_38_0_MIN,
+                IntegerValue(numeric38_0Min),
+                null,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION
+            ),
+            case(
+                NUMERIC_38_0_MAX_PLUS_ONE,
+                IntegerValue(numeric38_0Max + BigInteger.ONE),
+                null,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION
+            ),
+            case(
+                NUMERIC_38_0_MIN_MINUS_ONE,
+                IntegerValue(numeric38_0Min - BigInteger.ONE),
+                null,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION
+            ),
+        )
+
+    /**
+     * Many destination warehouses represent integers as a fixed-point type with 38 digits of
+     * precision. In this case, we only need to null out numbers larger than `1e38 - 1` / smaller
+     * than `-1e38 + 1`.
+     */
+    val numeric38_0 =
+        listOf(
+            case(NULL, NullValue, null),
+            case(ZERO, IntegerValue(0), bigint(0L)),
+            case(ONE, IntegerValue(1), bigint(1L)),
+            case(NEGATIVE_ONE, IntegerValue(-1), bigint(-1L)),
+            case(FORTY_TWO, IntegerValue(42), bigint(42L)),
+            case(NEGATIVE_FORTY_TWO, IntegerValue(-42), bigint(-42L)),
+            // int32 bounds, and slightly out of bounds
+            case(
+                INT32_MAX,
+                IntegerValue(Integer.MAX_VALUE.toLong()),
+                bigint(Integer.MAX_VALUE.toLong())
+            ),
+            case(
+                INT32_MIN,
+                IntegerValue(Integer.MIN_VALUE.toLong()),
+                bigint(Integer.MIN_VALUE.toLong())
+            ),
+            case(
+                INT32_MAX_PLUS_ONE,
+                IntegerValue(Integer.MAX_VALUE.toLong() + 1),
+                bigint(Integer.MAX_VALUE.toLong() + 1)
+            ),
+            case(
+                INT32_MIN_MINUS_ONE,
+                IntegerValue(Integer.MIN_VALUE.toLong() - 1),
+                bigint(Integer.MIN_VALUE.toLong() - 1)
+            ),
+            // int64 bounds, and slightly out of bounds
+            case(INT64_MAX, IntegerValue(Long.MAX_VALUE), bigint(Long.MAX_VALUE)),
+            case(INT64_MIN, IntegerValue(Long.MIN_VALUE), bigint(Long.MIN_VALUE)),
+            case(
+                INT64_MAX_PLUS_ONE,
+                IntegerValue(bigint(Long.MAX_VALUE) + BigInteger.ONE),
+                bigint(Long.MAX_VALUE) + BigInteger.ONE
+            ),
+            case(
+                INT64_MIN_MINUS_1,
+                IntegerValue(bigint(Long.MIN_VALUE) - BigInteger.ONE),
+                bigint(Long.MIN_VALUE) - BigInteger.ONE
+            ),
+            // NUMERIC(38, 9) bounds, and slightly out of bounds
+            case(NUMERIC_38_0_MAX, IntegerValue(numeric38_0Max), numeric38_0Max),
+            case(NUMERIC_38_0_MIN, IntegerValue(numeric38_0Min), numeric38_0Min),
+            // These values exceed the 38-digit range, so they get nulled out
+            case(
+                NUMERIC_38_0_MAX_PLUS_ONE,
+                IntegerValue(numeric38_0Max + BigInteger.ONE),
+                null,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION
+            ),
+            case(
+                NUMERIC_38_0_MIN_MINUS_ONE,
+                IntegerValue(numeric38_0Min - BigInteger.ONE),
+                null,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION
+            ),
+        )
+
+    @JvmStatic fun int64() = int64.toArgs()
+
+    /**
+     * Convenience fixture if your [TestTableOperationsClient] returns integers as [BigInteger]
+     * rather than [Long].
+     */
+    @JvmStatic
+    fun int64AsBigInteger() =
+        int64.map { it.copy(outputValue = it.outputValue?.let { bigint(it as Long) }) }
+
+    /**
+     * Convenience fixture if your [TestTableOperationsClient] returns integers as [BigDecimal]
+     * rather than [Long].
+     */
+    @JvmStatic
+    fun int64AsBigDecimal() =
+        int64.map { it.copy(outputValue = it.outputValue?.let { BigDecimal.valueOf(it as Long) }) }
+
+    @JvmStatic fun numeric38_0() = numeric38_0.toArgs()
+}
+
+object DataCoercionNumberFixtures {
+    val numeric38_9Max = bigdec("99999999999999999999999999999.999999999")
+    val numeric38_9Min = bigdec("-99999999999999999999999999999.999999999")
+
+    const val ZERO = "0"
+    const val ONE = "1"
+    const val NEGATIVE_ONE = "-1"
+    const val ONE_HUNDRED_TWENTY_THREE_POINT_FOUR = "123.4"
+    const val NEGATIVE_ONE_HUNDRED_TWENTY_THREE_POINT_FOUR = "123.4"
+    const val POSITIVE_HIGH_PRECISION_FLOAT = "positive high-precision float"
+    const val NEGATIVE_HIGH_PRECISION_FLOAT = "negative high-precision float"
+    const val NUMERIC_38_9_MAX = "numeric(38,9) max"
+    const val NUMERIC_38_9_MIN = "numeric(38,9) min"
+    const val SMALLEST_POSITIVE_FLOAT32 = "smallest positive float32"
+    const val SMALLEST_NEGATIVE_FLOAT32 = "smallest negative float32"
+    const val LARGEST_POSITIVE_FLOAT32 = "largest positive float32"
+    const val LARGEST_NEGATIVE_FLOAT32 = "largest negative float32"
+    const val SMALLEST_POSITIVE_FLOAT64 = "smallest positive float64"
+    const val SMALLEST_NEGATIVE_FLOAT64 = "smallest negative float64"
+    const val LARGEST_POSITIVE_FLOAT64 = "largest positive float64"
+    const val LARGEST_NEGATIVE_FLOAT64 = "largest negative float64"
+    const val SLIGHTLY_ABOVE_LARGEST_POSITIVE_FLOAT64 = "slightly above largest positive float64"
+    const val SLIGHTLY_BELOW_LARGEST_NEGATIVE_FLOAT64 = "slightly below largest negative float64"
+
+    val float64 =
+        listOf(
+            case(NULL, NullValue, null),
+            case(ZERO, NumberValue(bigdec(0)), 0.0),
+            case(ONE, NumberValue(bigdec(1)), 1.0),
+            case(NEGATIVE_ONE, NumberValue(bigdec(-1)), -1.0),
+            // This value isn't exactly representable as a float64
+            // (the exact value is `123.400000000000005684341886080801486968994140625`)
+            // but we should preserve the canonical representation
+            case(ONE_HUNDRED_TWENTY_THREE_POINT_FOUR, NumberValue(bigdec("123.4")), 123.4),
+            case(
+                NEGATIVE_ONE_HUNDRED_TWENTY_THREE_POINT_FOUR,
+                NumberValue(bigdec("-123.4")),
+                -123.4
+            ),
+            // These values have too much precision for a float64, so we round them
+            case(
+                POSITIVE_HIGH_PRECISION_FLOAT,
+                NumberValue(bigdec("1234567890.1234567890123456789")),
+                1234567890.1234567,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION
+            ),
+            case(
+                NEGATIVE_HIGH_PRECISION_FLOAT,
+                NumberValue(bigdec("-1234567890.1234567890123456789")),
+                -1234567890.1234567,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION
+            ),
+            case(
+                NUMERIC_38_9_MAX,
+                NumberValue(numeric38_9Max),
+                1.0E29,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION
+            ),
+            case(
+                NUMERIC_38_9_MIN,
+                NumberValue(numeric38_9Min),
+                -1.0E29,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION
+            ),
+            // min/max_value are all positive values, so we need to manually test their negative
+            // version
+            case(
+                SMALLEST_POSITIVE_FLOAT32,
+                NumberValue(bigdec(Float.MIN_VALUE.toDouble())),
+                Float.MIN_VALUE.toDouble()
+            ),
+            case(
+                SMALLEST_NEGATIVE_FLOAT32,
+                NumberValue(bigdec(-Float.MIN_VALUE.toDouble())),
+                -Float.MIN_VALUE.toDouble()
+            ),
+            case(
+                LARGEST_POSITIVE_FLOAT32,
+                NumberValue(bigdec(Float.MAX_VALUE.toDouble())),
+                Float.MAX_VALUE.toDouble()
+            ),
+            case(
+                LARGEST_NEGATIVE_FLOAT32,
+                NumberValue(bigdec(-Float.MAX_VALUE.toDouble())),
+                -Float.MAX_VALUE.toDouble()
+            ),
+            case(
+                SMALLEST_POSITIVE_FLOAT64,
+                NumberValue(bigdec(Double.MIN_VALUE)),
+                Double.MIN_VALUE
+            ),
+            case(
+                SMALLEST_NEGATIVE_FLOAT64,
+                NumberValue(bigdec(-Double.MIN_VALUE)),
+                -Double.MIN_VALUE
+            ),
+            case(LARGEST_POSITIVE_FLOAT64, NumberValue(bigdec(Double.MAX_VALUE)), Double.MAX_VALUE),
+            case(
+                LARGEST_NEGATIVE_FLOAT64,
+                NumberValue(bigdec(-Double.MAX_VALUE)),
+                -Double.MAX_VALUE
+            ),
+            // These values are out of bounds, so we null them
+            case(
+                SLIGHTLY_ABOVE_LARGEST_POSITIVE_FLOAT64,
+                NumberValue(bigdec(Double.MAX_VALUE) + bigdec(Double.MIN_VALUE)),
+                null,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION
+            ),
+            case(
+                SLIGHTLY_BELOW_LARGEST_NEGATIVE_FLOAT64,
+                NumberValue(bigdec(-Double.MAX_VALUE) - bigdec(Double.MIN_VALUE)),
+                null,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION
+            ),
+        )
+
+    val numeric38_9 =
+        listOf(
+                case(NULL, NullValue, null),
+                case(ZERO, NumberValue(bigdec(0)), bigdec(0.0)),
+                case(ONE, NumberValue(bigdec(1)), bigdec(1.0)),
+                case(NEGATIVE_ONE, NumberValue(bigdec(-1)), bigdec(-1.0)),
+                // This value isn't exactly representable as a float64
+                // (the exact value is `123.400000000000005684341886080801486968994140625`)
+                // but it's perfectly fine as a numeric(38, 9)
+                case(
+                    ONE_HUNDRED_TWENTY_THREE_POINT_FOUR,
+                    NumberValue(bigdec("123.4")),
+                    bigdec("123.4")
+                ),
+                case(
+                    NEGATIVE_ONE_HUNDRED_TWENTY_THREE_POINT_FOUR,
+                    NumberValue(bigdec("-123.4")),
+                    bigdec("-123.4")
+                ),
+                // These values have too much precision for a numeric(38, 9), so we round them
+                case(
+                    POSITIVE_HIGH_PRECISION_FLOAT,
+                    NumberValue(bigdec("1234567890.1234567890123456789")),
+                    bigdec("1234567890.123456789"),
+                    Reason.DESTINATION_FIELD_SIZE_LIMITATION
+                ),
+                case(
+                    NEGATIVE_HIGH_PRECISION_FLOAT,
+                    NumberValue(bigdec("-1234567890.1234567890123456789")),
+                    bigdec("-1234567890.123456789"),
+                    Reason.DESTINATION_FIELD_SIZE_LIMITATION
+                ),
+                case(
+                    SMALLEST_POSITIVE_FLOAT32,
+                    NumberValue(bigdec(Float.MIN_VALUE.toDouble())),
+                    bigdec(0),
+                    Reason.DESTINATION_FIELD_SIZE_LIMITATION
+                ),
+                case(
+                    SMALLEST_NEGATIVE_FLOAT32,
+                    NumberValue(bigdec(-Float.MIN_VALUE.toDouble())),
+                    bigdec(0),
+                    Reason.DESTINATION_FIELD_SIZE_LIMITATION
+                ),
+                case(
+                    SMALLEST_POSITIVE_FLOAT64,
+                    NumberValue(bigdec(Double.MIN_VALUE)),
+                    bigdec(0),
+                    Reason.DESTINATION_FIELD_SIZE_LIMITATION
+                ),
+                case(
+                    SMALLEST_NEGATIVE_FLOAT64,
+                    NumberValue(bigdec(-Double.MIN_VALUE)),
+                    bigdec(0),
+                    Reason.DESTINATION_FIELD_SIZE_LIMITATION
+                ),
+                // numeric bounds are perfectly fine
+                case(NUMERIC_38_9_MAX, NumberValue(numeric38_9Max), numeric38_9Max),
+                case(NUMERIC_38_9_MIN, NumberValue(numeric38_9Min), numeric38_9Min),
+                // These values are out of bounds, so we null them
+                case(
+                    LARGEST_POSITIVE_FLOAT32,
+                    NumberValue(bigdec(Float.MAX_VALUE.toDouble())),
+                    null,
+                    Reason.DESTINATION_FIELD_SIZE_LIMITATION
+                ),
+                case(
+                    LARGEST_NEGATIVE_FLOAT32,
+                    NumberValue(bigdec(-Float.MAX_VALUE.toDouble())),
+                    null,
+                    Reason.DESTINATION_FIELD_SIZE_LIMITATION
+                ),
+                case(
+                    LARGEST_POSITIVE_FLOAT64,
+                    NumberValue(bigdec(Double.MAX_VALUE)),
+                    null,
+                    Reason.DESTINATION_FIELD_SIZE_LIMITATION
+                ),
+                case(
+                    LARGEST_NEGATIVE_FLOAT64,
+                    NumberValue(bigdec(-Double.MAX_VALUE)),
+                    null,
+                    Reason.DESTINATION_FIELD_SIZE_LIMITATION
+                ),
+                case(
+                    SLIGHTLY_ABOVE_LARGEST_POSITIVE_FLOAT64,
+                    NumberValue(bigdec(Double.MAX_VALUE) + bigdec(Double.MIN_VALUE)),
+                    null,
+                    Reason.DESTINATION_FIELD_SIZE_LIMITATION
+                ),
+                case(
+                    SLIGHTLY_BELOW_LARGEST_NEGATIVE_FLOAT64,
+                    NumberValue(bigdec(-Double.MAX_VALUE) - bigdec(Double.MIN_VALUE)),
+                    null,
+                    Reason.DESTINATION_FIELD_SIZE_LIMITATION
+                ),
+            )
+            .map { it.copy(outputValue = (it.outputValue as BigDecimal?)?.setScale(9)) }
+
+    @JvmStatic fun float64() = float64.toArgs()
+    @JvmStatic fun numeric38_9() = numeric38_9.toArgs()
+}
+
+const val SIMPLE_TIMESTAMP = "simple timestamp"
+const val UNIX_EPOCH = "unix epoch"
+const val MINIMUM_TIMESTAMP = "minimum timestamp"
+const val MAXIMUM_TIMESTAMP = "maximum timestamp"
+const val OUT_OF_RANGE_TIMESTAMP = "out of range timestamp"
+const val HIGH_PRECISION_TIMESTAMP = "high-precision timestamp"
+
+object DataCoercionTimestampTzFixtures {
+    /**
+     * Many warehouses support timestamps between years 0001 - 9999.
+     *
+     * Depending on the exact warehouse, you may need to tweak the precision on some values. For
+     * example, Snowflake supports nanoseconds-precision timestamps (9 decimal points), but Bigquery
+     * only supports microseconds-precision (6 decimal points). Bigquery would probably do something
+     * like:
+     * ```kotlin
+     * DataCoercionNumberFixtures.traditionalWarehouse
+     *   .map {
+     *     when (it.name) {
+     *       "maximum AD timestamp" -> it.copy(
+     *         inputValue = TimestampWithTimezoneValue("9999-12-31T23:59:59.999999Z"),
+     *         outputValue = OffsetDateTime.parse("9999-12-31T23:59:59.999999Z"),
+     *         changeReason = Reason.DESTINATION_FIELD_SIZE_LIMITATION,
+     *       )
+     *       "high-precision timestamp" -> it.copy(
+     *         outputValue = OffsetDateTime.parse("2025-01-23T01:01:00.123456Z"),
+     *         changeReason = Reason.DESTINATION_FIELD_SIZE_LIMITATION,
+     *       )
+     *     }
+     *   }
+     * ```
+     */
+    val commonWarehouse =
+        listOf(
+            case(NULL, NullValue, null),
+            case(
+                SIMPLE_TIMESTAMP,
+                TimestampWithTimezoneValue("2025-01-23T12:34:56.789Z"),
+                "2025-01-23T12:34:56.789Z",
+            ),
+            case(
+                UNIX_EPOCH,
+                TimestampWithTimezoneValue("1970-01-01T00:00:00Z"),
+                "1970-01-01T00:00:00Z",
+            ),
+            case(
+                MINIMUM_TIMESTAMP,
+                TimestampWithTimezoneValue("0001-01-01T00:00:00Z"),
+                "0001-01-01T00:00:00Z",
+            ),
+            case(
+                MAXIMUM_TIMESTAMP,
+                TimestampWithTimezoneValue("9999-12-31T23:59:59.999999999Z"),
+                "9999-12-31T23:59:59.999999999Z",
+            ),
+            case(
+                OUT_OF_RANGE_TIMESTAMP,
+                TimestampWithTimezoneValue(odt("10000-01-01T00:00Z")),
+                null,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION,
+            ),
+            case(
+                HIGH_PRECISION_TIMESTAMP,
+                TimestampWithTimezoneValue("2025-01-23T01:01:00.123456789Z"),
+                "2025-01-23T01:01:00.123456789Z",
+            ),
+        )
+
+    @JvmStatic fun commonWarehouse() = commonWarehouse.toArgs()
+}
+
+object DataCoercionTimestampNtzFixtures {
+    /** See [DataCoercionTimestampTzFixtures.commonWarehouse] for explanation */
+    val commonWarehouse =
+        listOf(
+            case(NULL, NullValue, null),
+            case(
+                SIMPLE_TIMESTAMP,
+                TimestampWithoutTimezoneValue("2025-01-23T12:34:56.789"),
+                "2025-01-23T12:34:56.789",
+            ),
+            case(
+                UNIX_EPOCH,
+                TimestampWithoutTimezoneValue("1970-01-01T00:00:00"),
+                "1970-01-01T00:00:00",
+            ),
+            case(
+                MINIMUM_TIMESTAMP,
+                TimestampWithoutTimezoneValue("0001-01-01T00:00:00"),
+                "0001-01-01T00:00:00",
+            ),
+            case(
+                MAXIMUM_TIMESTAMP,
+                TimestampWithoutTimezoneValue("9999-12-31T23:59:59.999999999"),
+                "9999-12-31T23:59:59.999999999",
+            ),
+            case(
+                OUT_OF_RANGE_TIMESTAMP,
+                TimestampWithoutTimezoneValue(ldt("10000-01-01T00:00")),
+                null,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION,
+            ),
+            case(
+                HIGH_PRECISION_TIMESTAMP,
+                TimestampWithoutTimezoneValue("2025-01-23T01:01:00.123456789"),
+                "2025-01-23T01:01:00.123456789",
+            ),
+        )
+
+    @JvmStatic fun commonWarehouse() = commonWarehouse.toArgs()
+}
+
+const val MIDNIGHT = "midnight"
+const val MAX_TIME = "max time"
+const val HIGH_NOON = "high noon"
+
+object DataCoercionTimeTzFixtures {
+    val timetz =
+        listOf(
+            case(NULL, NullValue, null),
+            case(MIDNIGHT, TimeWithTimezoneValue("00:00Z"), "00:00Z"),
+            case(MAX_TIME, TimeWithTimezoneValue("23:59:59.999999999Z"), "23:59:59.999999999Z"),
+            case(HIGH_NOON, TimeWithTimezoneValue("12:00Z"), "12:00Z"),
+        )
+
+    @JvmStatic fun timetz() = timetz.toArgs()
+}
+
+object DataCoercionTimeNtzFixtures {
+    val timentz =
+        listOf(
+            case(NULL, NullValue, null),
+            case(MIDNIGHT, TimeWithoutTimezoneValue("00:00"), "00:00"),
+            case(MAX_TIME, TimeWithoutTimezoneValue("23:59:59.999999999"), "23:59:59.999999999"),
+            case(HIGH_NOON, TimeWithoutTimezoneValue("12:00"), "12:00"),
+        )
+
+    @JvmStatic fun timentz() = timentz.toArgs()
+}
+
+object DataCoercionDateFixtures {
+    val commonWarehouse =
+        listOf(
+            case(NULL, NullValue, null),
+            case(
+                SIMPLE_TIMESTAMP,
+                DateValue("2025-01-23"),
+                "2025-01-23",
+            ),
+            case(
+                UNIX_EPOCH,
+                DateValue("1970-01-01"),
+                "1970-01-01",
+            ),
+            case(
+                MINIMUM_TIMESTAMP,
+                DateValue("0001-01-01"),
+                "0001-01-01",
+            ),
+            case(
+                MAXIMUM_TIMESTAMP,
+                DateValue("9999-12-31"),
+                "9999-12-31",
+            ),
+            case(
+                OUT_OF_RANGE_TIMESTAMP,
+                DateValue(date("10000-01-01")),
+                null,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION,
+            ),
+        )
+
+    @JvmStatic fun commonWarehouse() = commonWarehouse.toArgs()
+}
+
+object DataCoercionStringFixtures {
+    const val EMPTY_STRING = "empty string"
+    const val SHORT_STRING = "short string"
+    const val LONG_STRING = "long string"
+    const val SPECIAL_CHARS_STRING = "special chars string"
+
+    val strings =
+        listOf(
+            case(NULL, NullValue, null),
+            case(EMPTY_STRING, StringValue(""), ""),
+            case(SHORT_STRING, StringValue("foo"), "foo"),
+            // Implementers may override this to test their destination-specific limits.
+            // The default value is 8MB + 1 byte (slightly longer than snowflake's varchar limit).
+            case(
+                LONG_STRING,
+                StringValue("a".repeat(16777216 + 1)),
+                null,
+                Reason.DESTINATION_FIELD_SIZE_LIMITATION
+            ),
+            case(
+                SPECIAL_CHARS_STRING,
+                StringValue("`~!@#$%^&*()-=_+[]\\{}|o'O\",./<>?)Δ⅀↑∀"),
+                "`~!@#$%^&*()-=_+[]\\{}|o'O\",./<>?)Δ⅀↑∀"
+            ),
+        )
+
+    @JvmStatic fun strings() = strings.toArgs()
+}
+
+object DataCoercionObjectFixtures {
+    const val EMPTY_OBJECT = "empty object"
+    const val NORMAL_OBJECT = "normal object"
+
+    val objects =
+        listOf(
+            case(NULL, NullValue, null),
+            case(EMPTY_OBJECT, ObjectValue(linkedMapOf()), emptyMap<String, Any?>()),
+            case(
+                NORMAL_OBJECT,
+                ObjectValue(linkedMapOf("foo" to StringValue("bar"))),
+                mapOf("foo" to "bar")
+            ),
+        )
+
+    val stringifiedObjects =
+        objects.map { fixture ->
+            fixture.copy(outputValue = fixture.outputValue?.serializeToString())
+        }
+
+    @JvmStatic fun objects() = objects.toArgs()
+
+    @JvmStatic fun stringifiedObjects() = stringifiedObjects.toArgs()
+}
+
+object DataCoercionArrayFixtures {
+    const val EMPTY_ARRAY = "empty array"
+    const val NORMAL_ARRAY = "normal array"
+
+    val arrays =
+        listOf(
+            case(NULL, NullValue, null),
+            case(EMPTY_ARRAY, ArrayValue(emptyList()), emptyList<Any?>()),
+            case(NORMAL_ARRAY, ArrayValue(listOf(StringValue("foo"))), listOf("foo")),
+        )
+
+    val stringifiedArrays =
+        arrays.map { fixture ->
+            fixture.copy(outputValue = fixture.outputValue?.serializeToString())
+        }
+
+    @JvmStatic fun arrays() = arrays.toArgs()
+
+    @JvmStatic fun stringifiedArrays() = stringifiedArrays.toArgs()
+}
+
+const val UNION_INT_VALUE = "int value"
+const val UNION_OBJ_VALUE = "object value"
+const val UNION_STR_VALUE = "string value"
+
+object DataCoercionUnionFixtures {
+    val unions =
+        listOf(
+            case(NULL, NullValue, null),
+            case(UNION_INT_VALUE, IntegerValue(42), 42L),
+            case(UNION_STR_VALUE, StringValue("foo"), "foo"),
+            case(
+                UNION_OBJ_VALUE,
+                ObjectValue(linkedMapOf("foo" to StringValue("bar"))),
+                mapOf("foo" to "bar")
+            ),
+        )
+
+    val stringifiedUnions =
+        unions.map { fixture ->
+            fixture.copy(outputValue = fixture.outputValue?.serializeToString())
+        }
+
+    @JvmStatic fun unions() = unions.toArgs()
+
+    @JvmStatic fun stringifiedUnions() = stringifiedUnions.toArgs()
+}
+
+object DataCoercionLegacyUnionFixtures {
+    val unions =
+        listOf(
+            case(NULL, NullValue, null),
+            // Legacy union of int x object will select object, and you can't write an int to an
+            // object column.
+            // So we should null it out.
+            case(UNION_INT_VALUE, IntegerValue(42), null, Reason.DESTINATION_TYPECAST_ERROR),
+            // Similarly, we should null out strings.
+            case(UNION_STR_VALUE, StringValue("foo"), "foo"),
+            // But objects can be written as objects, so retain this value.
+            case(
+                UNION_OBJ_VALUE,
+                ObjectValue(linkedMapOf("foo" to StringValue("bar"))),
+                mapOf("foo" to "bar")
+            ),
+        )
+
+    val stringifiedUnions =
+        DataCoercionUnionFixtures.unions.map { fixture ->
+            fixture.copy(outputValue = fixture.outputValue?.serializeToString())
+        }
+
+    @JvmStatic fun unions() = unions.toArgs()
+
+    @JvmStatic fun stringifiedUnions() = DataCoercionUnionFixtures.stringifiedUnions.toArgs()
+}
+
+// This is pretty much identical to UnionFixtures, but separating them in case we need to add
+// different test cases for either of them.
+object DataCoercionUnknownFixtures {
+    const val INT_VALUE = "integer value"
+    const val STR_VALUE = "string value"
+    const val OBJ_VALUE = "object value"
+
+    val unknowns =
+        listOf(
+            case(NULL, NullValue, null),
+            case(INT_VALUE, IntegerValue(42), 42L),
+            case(STR_VALUE, StringValue("foo"), "foo"),
+            case(
+                OBJ_VALUE,
+                ObjectValue(linkedMapOf("foo" to StringValue("bar"))),
+                mapOf("foo" to "bar")
+            ),
+        )
+
+    val stringifiedUnknowns =
+        unknowns.map { fixture ->
+            fixture.copy(outputValue = fixture.outputValue?.serializeToString())
+        }
+
+    @JvmStatic fun unknowns() = unknowns.toArgs()
+
+    @JvmStatic fun stringifiedUnknowns() = stringifiedUnknowns.toArgs()
+}
+
+fun List<DataCoercionTestCase>.toArgs(): List<Arguments> =
+    this.map { Arguments.argumentSet(it.name, it.inputValue, it.outputValue, it.changeReason) }
+        .toList()
+
+/**
+ * Utility method to use the BigDecimal constructor (supports exponential notation like `1e38`) to
+ * construct a BigInteger.
+ */
+fun bigint(str: String): BigInteger = BigDecimal(str).toBigIntegerExact()
+
+/** Shorthand utility method to construct a bigint from a long */
+fun bigint(long: Long): BigInteger = BigInteger.valueOf(long)
+
+fun bigdec(str: String): BigDecimal = BigDecimal(str)
+
+fun bigdec(double: Double): BigDecimal = BigDecimal.valueOf(double)
+
+fun bigdec(int: Int): BigDecimal = BigDecimal.valueOf(int.toDouble())
+
+fun odt(str: String): OffsetDateTime = OffsetDateTime.parse(str, dateTimeFormatter)
+
+fun ldt(str: String): LocalDateTime = LocalDateTime.parse(str, dateTimeFormatter)
+
+fun date(str: String): LocalDate = LocalDate.parse(str, dateFormatter)
+
+// The default java.time.*.parse() behavior only accepts up to 4-digit years.
+// Build a custom formatter to handle larger years.
+val dateFormatter =
+    DateTimeFormatterBuilder()
+        // java.time.* supports up to 9-digit years
+        .appendValue(ChronoField.YEAR, 1, 9, SignStyle.NORMAL)
+        .appendLiteral('-')
+        .appendValue(ChronoField.MONTH_OF_YEAR)
+        .appendLiteral('-')
+        .appendValue(ChronoField.DAY_OF_MONTH)
+        .toFormatter()
+
+val dateTimeFormatter =
+    DateTimeFormatterBuilder()
+        .append(dateFormatter)
+        .appendLiteral('T')
+        // Accepts strings with/without an offset, so we can use this formatter
+        // for both timestamp with and without timezone
+        .append(DateTimeFormatter.ISO_TIME)
+        .toFormatter()
+
+/**
+ * Represents a single data coercion test case. You probably want to use [case] as a shorthand
+ * constructor.
+ *
+ * @param name A short human-readable name for the test. Primarily useful for tests where
+ * [inputValue] is either very long, or otherwise hard to read.
+ * @param inputValue The value to pass into [ValueCoercer.validate]
+ * @param outputValue The value that we expect to read back from the destination. Should be
+ * basically equivalent to the output of [ValueCoercer.validate]
+ * @param changeReason If `validate` returns Truncate/Nullify, the reason for that
+ * truncation/nullification. If `validate` returns Valid, this should be null.
+ */
+data class DataCoercionTestCase(
+    val name: String,
+    val inputValue: AirbyteValue,
+    val outputValue: Any?,
+    val changeReason: Reason? = null,
+)
+
+fun case(
+    name: String,
+    inputValue: AirbyteValue,
+    outputValue: Any?,
+    changeReason: Reason? = null,
+) = DataCoercionTestCase(name, inputValue, outputValue, changeReason)
+
+const val NULL = "null"

airbyte-cdk/bulk/core/load/src/testFixtures/kotlin/io/airbyte/cdk/load/component/DataCoercionSuite.kt

@@ -0,0 +1,369 @@
+/*
+ * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
+ */
+
+package io.airbyte.cdk.load.component
+
+import io.airbyte.cdk.load.data.AirbyteValue
+import io.airbyte.cdk.load.data.ArrayType
+import io.airbyte.cdk.load.data.ArrayTypeWithoutSchema
+import io.airbyte.cdk.load.data.BooleanType
+import io.airbyte.cdk.load.data.BooleanValue
+import io.airbyte.cdk.load.data.DateType
+import io.airbyte.cdk.load.data.FieldType
+import io.airbyte.cdk.load.data.IntegerType
+import io.airbyte.cdk.load.data.NumberType
+import io.airbyte.cdk.load.data.ObjectType
+import io.airbyte.cdk.load.data.ObjectTypeWithEmptySchema
+import io.airbyte.cdk.load.data.ObjectTypeWithoutSchema
+import io.airbyte.cdk.load.data.StringType
+import io.airbyte.cdk.load.data.TimeTypeWithTimezone
+import io.airbyte.cdk.load.data.TimeTypeWithoutTimezone
+import io.airbyte.cdk.load.data.TimestampTypeWithTimezone
+import io.airbyte.cdk.load.data.TimestampTypeWithoutTimezone
+import io.airbyte.cdk.load.data.UnionType
+import io.airbyte.cdk.load.data.UnknownType
+import io.airbyte.cdk.load.dataflow.transform.ValueCoercer
+import io.airbyte.cdk.load.message.Meta
+import io.airbyte.cdk.load.schema.TableSchemaFactory
+import io.airbyte.cdk.load.table.ColumnNameMapping
+import io.airbyte.cdk.load.util.Jsons
+import io.airbyte.protocol.models.v0.AirbyteRecordMessageMetaChange.Reason
+import io.micronaut.test.extensions.junit5.annotation.MicronautTest
+import kotlinx.coroutines.test.runTest
+
+/**
+ * The tests in this class are designed to reference the parameters defined in
+ * `DataCoercionFixtures.kt`. For example, you might annotate [`handle integer values`] with
+ * `@MethodSource("io.airbyte.cdk.load.component.DataCoercionIntegerFixtures#int32")`. See each
+ * fixture class for explanations of what behavior they are exercising.
+ *
+ * Note that this class _only_ exercises [ValueCoercer.validate]. You should write separate unit
+ * tests for [ValueCoercer.map]. For now, the `map` function is primarily intended for transforming
+ * `UnionType` fields into other types (typically `StringType`), at which point your `validate`
+ * implementation should be able to handle any StringValue (regardless of whether it was originally
+ * a StringType or UnionType).
+ */
+@MicronautTest(environments = ["component"], resolveParameters = false)
+interface DataCoercionSuite {
+    val coercer: ValueCoercer
+    val airbyteMetaColumnMapping: Map<String, String>
+        get() = Meta.COLUMN_NAMES.associateWith { it }
+    val columnNameMapping: ColumnNameMapping
+        get() = ColumnNameMapping(mapOf("test" to "test"))
+
+    val opsClient: TableOperationsClient
+    val testClient: TestTableOperationsClient
+    val schemaFactory: TableSchemaFactory
+
+    val harness: TableOperationsTestHarness
+        get() =
+            TableOperationsTestHarness(
+                opsClient,
+                testClient,
+                schemaFactory,
+                airbyteMetaColumnMapping
+            )
+
+    /** Fixtures are defined in [DataCoercionIntegerFixtures]. */
+    fun `handle integer values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(IntegerType, nullable = true),
+            inputValue,
+            expectedValue,
+            expectedChangeReason,
+        )
+    }
+
+    /** Fixtures are defined in [DataCoercionNumberFixtures]. */
+    fun `handle number values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(NumberType, nullable = true),
+            inputValue,
+            expectedValue,
+            expectedChangeReason,
+        )
+    }
+
+    /** Fixtures are defined in [DataCoercionTimestampTzFixtures]. */
+    fun `handle timestamptz values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(TimestampTypeWithTimezone, nullable = true),
+            inputValue,
+            expectedValue,
+            expectedChangeReason,
+        )
+    }
+
+    /** Fixtures are defined in [DataCoercionTimestampNtzFixtures]. */
+    fun `handle timestampntz values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(TimestampTypeWithoutTimezone, nullable = true),
+            inputValue,
+            expectedValue,
+            expectedChangeReason,
+        )
+    }
+
+    /** Fixtures are defined in [DataCoercionTimeTzFixtures]. */
+    fun `handle timetz values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(TimeTypeWithTimezone, nullable = true),
+            inputValue,
+            expectedValue,
+            expectedChangeReason,
+        )
+    }
+
+    /** Fixtures are defined in [DataCoercionTimeNtzFixtures]. */
+    fun `handle timentz values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(TimeTypeWithoutTimezone, nullable = true),
+            inputValue,
+            expectedValue,
+            expectedChangeReason,
+        )
+    }
+
+    /** Fixtures are defined in [DataCoercionDateFixtures]. */
+    fun `handle date values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(DateType, nullable = true),
+            inputValue,
+            expectedValue,
+            expectedChangeReason,
+        )
+    }
+
+    /** No fixtures, hardcoded to just write `true` */
+    fun `handle bool values`(expectedValue: Any?) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(BooleanType, nullable = true),
+            // Just test on `true` and assume `false` also works
+            BooleanValue(true),
+            expectedValue,
+            // If your destination is nulling/truncating booleans... that's almost definitely a bug
+            expectedChangeReason = null,
+        )
+    }
+
+    /** Fixtures are defined in [DataCoercionStringFixtures]. */
+    fun `handle string values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(StringType, nullable = true),
+            inputValue,
+            expectedValue,
+            expectedChangeReason,
+        )
+    }
+
+    /** Fixtures are defined in [DataCoercionObjectFixtures]. */
+    fun `handle object values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(
+                ObjectType(linkedMapOf("foo" to FieldType(StringType, true))),
+                nullable = true
+            ),
+            inputValue,
+            expectedValue,
+            expectedChangeReason,
+        )
+    }
+
+    /** Fixtures are defined in [DataCoercionObjectFixtures]. */
+    fun `handle empty object values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(ObjectTypeWithEmptySchema, nullable = true),
+            inputValue,
+            expectedValue,
+            expectedChangeReason,
+        )
+    }
+
+    /** Fixtures are defined in [DataCoercionObjectFixtures]. */
+    fun `handle schemaless object values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(ObjectTypeWithoutSchema, nullable = true),
+            inputValue,
+            expectedValue,
+            expectedChangeReason,
+        )
+    }
+
+    /** Fixtures are defined in [DataCoercionArrayFixtures]. */
+    fun `handle array values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(ArrayType(FieldType(StringType, true)), nullable = true),
+            inputValue,
+            expectedValue,
+            expectedChangeReason,
+        )
+    }
+
+    /** Fixtures are defined in [DataCoercionArrayFixtures]. */
+    fun `handle schemaless array values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(ArrayTypeWithoutSchema, nullable = true),
+            inputValue,
+            expectedValue,
+            expectedChangeReason,
+        )
+    }
+
+    /**
+     * All destinations should implement this, even if your destination is supporting legacy unions.
+     *
+     * Fixtures are defined in [DataCoercionUnionFixtures].
+     */
+    fun `handle union values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(
+                UnionType(
+                    setOf(
+                        ObjectType(linkedMapOf("foo" to FieldType(StringType, true))),
+                        IntegerType,
+                        StringType,
+                    ),
+                    isLegacyUnion = false
+                ),
+                nullable = true
+            ),
+            inputValue,
+            expectedValue,
+            expectedChangeReason,
+        )
+    }
+
+    /**
+     * Only legacy destinations that are maintaining "legacy" union behavior should implement this
+     * test. If you're not sure, check whether your `application-connector.yaml` includes a
+     * `airbyte.destination.core.types.unions: LEGACY` property.
+     *
+     * Fixtures are defined in [DataCoercionLegacyUnionFixtures].
+     */
+    fun `handle legacy union values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(
+                UnionType(
+                    setOf(
+                        ObjectType(linkedMapOf("foo" to FieldType(StringType, true))),
+                        IntegerType,
+                        StringType,
+                    ),
+                    isLegacyUnion = true
+                ),
+                nullable = true
+            ),
+            inputValue,
+            expectedValue,
+            expectedChangeReason,
+        )
+    }
+
+    fun `handle unknown values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) = runTest {
+        harness.testValueCoercion(
+            coercer,
+            columnNameMapping,
+            FieldType(UnknownType(Jsons.readTree(("""{"type": "potato"}"""))), nullable = true),
+            inputValue,
+            expectedValue,
+            expectedChangeReason,
+        )
+    }
+}

airbyte-cdk/bulk/core/load/src/testFixtures/kotlin/io/airbyte/cdk/load/component/TableOperationsFixtures.kt

@@ -23,6 +23,7 @@ import io.airbyte.cdk.load.data.TimeTypeWithoutTimezone
 import io.airbyte.cdk.load.data.TimestampTypeWithTimezone
 import io.airbyte.cdk.load.data.TimestampTypeWithoutTimezone
 import io.airbyte.cdk.load.data.TimestampWithTimezoneValue
+import io.airbyte.cdk.load.data.UnionType
 import io.airbyte.cdk.load.data.UnknownType
 import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_EXTRACTED_AT
 import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_GENERATION_ID
@@ -84,6 +85,18 @@ object TableOperationsFixtures {
                 "array" to FieldType(ArrayType(FieldType(StringType, true)), true),
                 "object" to
                     FieldType(ObjectType(linkedMapOf("key" to FieldType(StringType, true))), true),
+                "union" to
+                    FieldType(
+                        UnionType(setOf(StringType, IntegerType), isLegacyUnion = false),
+                        true
+                    ),
+                // Most destinations just ignore the isLegacyUnion flag, which is totally fine.
+                // This is here for the small set of connectors that respect it.
+                "legacy_union" to
+                    FieldType(
+                        UnionType(setOf(StringType, IntegerType), isLegacyUnion = true),
+                        true
+                    ),
                 "unknown" to FieldType(UnknownType(Jsons.readTree("""{"type": "potato"}""")), true),
             ),
         )
@@ -101,6 +114,8 @@ object TableOperationsFixtures {
                 "time_ntz" to "time_ntz",
                 "array" to "array",
                 "object" to "object",
+                "union" to "union",
+                "legacy_union" to "legacy_union",
                 "unknown" to "unknown",
             )
         )
@@ -714,6 +729,11 @@ object TableOperationsFixtures {
         return map { record -> record.mapKeys { (k, _) -> totalMapping.invert()[k] ?: k } }
     }
 
+    fun <V> List<Map<String, V>>.removeAirbyteColumns(
+        airbyteMetaColumnMapping: Map<String, String>
+    ): List<Map<String, V>> =
+        this.map { rec -> rec.filter { !airbyteMetaColumnMapping.containsValue(it.key) } }
+
     fun <V> List<Map<String, V>>.removeNulls() =
         this.map { record -> record.filterValues { it != null } }
 

airbyte-cdk/bulk/core/load/src/testFixtures/kotlin/io/airbyte/cdk/load/component/TableOperationsSuite.kt

@@ -58,7 +58,8 @@ interface TableOperationsSuite {
         get() = Meta.COLUMN_NAMES.associateWith { it }
 
     private val harness: TableOperationsTestHarness
-        get() = TableOperationsTestHarness(client, testClient, airbyteMetaColumnMapping)
+        get() =
+            TableOperationsTestHarness(client, testClient, schemaFactory, airbyteMetaColumnMapping)
 
     /** Tests basic database connectivity by pinging the database. */
     fun `connect to database`() = runTest { assertDoesNotThrow { testClient.ping() } }
@@ -606,7 +607,7 @@ interface TableOperationsSuite {
         val targetTableSchema =
             schemaFactory.make(
                 targetTable,
-                Fixtures.TEST_INTEGER_SCHEMA.properties,
+                Fixtures.ID_TEST_WITH_CDC_SCHEMA.properties,
                 Dedupe(
                     primaryKey = listOf(listOf(Fixtures.ID_FIELD)),
                     cursor = listOf(Fixtures.TEST_FIELD),

airbyte-cdk/bulk/core/load/src/testFixtures/kotlin/io/airbyte/cdk/load/component/TableOperationsTestHarness.kt

@@ -4,11 +4,24 @@
 
 package io.airbyte.cdk.load.component
 
+import io.airbyte.cdk.load.command.Append
 import io.airbyte.cdk.load.command.DestinationStream
+import io.airbyte.cdk.load.component.TableOperationsFixtures.inputRecord
 import io.airbyte.cdk.load.component.TableOperationsFixtures.insertRecords
+import io.airbyte.cdk.load.component.TableOperationsFixtures.removeAirbyteColumns
+import io.airbyte.cdk.load.component.TableOperationsFixtures.removeNulls
+import io.airbyte.cdk.load.component.TableOperationsFixtures.reverseColumnNameMapping
 import io.airbyte.cdk.load.data.AirbyteValue
+import io.airbyte.cdk.load.data.EnrichedAirbyteValue
+import io.airbyte.cdk.load.data.FieldType
+import io.airbyte.cdk.load.data.NullValue
+import io.airbyte.cdk.load.data.ObjectType
+import io.airbyte.cdk.load.dataflow.transform.ValidationResult
+import io.airbyte.cdk.load.dataflow.transform.ValueCoercer
+import io.airbyte.cdk.load.schema.TableSchemaFactory
 import io.airbyte.cdk.load.schema.model.TableName
 import io.airbyte.cdk.load.table.ColumnNameMapping
+import io.airbyte.protocol.models.v0.AirbyteRecordMessageMetaChange.Reason
 import io.github.oshai.kotlinlogging.KotlinLogging
 import org.junit.jupiter.api.Assertions.assertEquals
 
@@ -21,6 +34,7 @@ private val log = KotlinLogging.logger {}
 class TableOperationsTestHarness(
     private val client: TableOperationsClient,
     private val testClient: TestTableOperationsClient,
+    private val schemaFactory: TableSchemaFactory,
     private val airbyteMetaColumnMapping: Map<String, String>,
 ) {
 
@@ -100,8 +114,77 @@ class TableOperationsTestHarness(
     /** Reads records from a table, filtering out Meta columns. */
     suspend fun readTableWithoutMetaColumns(tableName: TableName): List<Map<String, Any>> {
         val tableRead = testClient.readTable(tableName)
-        return tableRead.map { rec ->
-            rec.filter { !airbyteMetaColumnMapping.containsValue(it.key) }
+        return tableRead.removeAirbyteColumns(airbyteMetaColumnMapping)
+    }
+
+    /** Apply the coercer to a value and verify that we can write the coerced value correctly */
+    suspend fun testValueCoercion(
+        coercer: ValueCoercer,
+        columnNameMapping: ColumnNameMapping,
+        fieldType: FieldType,
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?,
+    ) {
+        val testNamespace = TableOperationsFixtures.generateTestNamespace("test")
+        val tableName =
+            TableOperationsFixtures.generateTestTableName("table-test-table", testNamespace)
+        val schema = ObjectType(linkedMapOf("test" to fieldType))
+        val tableSchema = schemaFactory.make(tableName, schema.properties, Append)
+        val stream =
+            TableOperationsFixtures.createStream(
+                namespace = tableName.namespace,
+                name = tableName.name,
+                tableSchema = tableSchema,
+            )
+
+        val inputValueAsEnrichedAirbyteValue =
+            EnrichedAirbyteValue(
+                inputValue,
+                fieldType.type,
+                "test",
+                airbyteMetaField = null,
+            )
+        val validatedValue = coercer.validate(inputValueAsEnrichedAirbyteValue)
+        val valueToInsert: AirbyteValue
+        val changeReason: Reason?
+        when (validatedValue) {
+            is ValidationResult.ShouldNullify -> {
+                valueToInsert = NullValue
+                changeReason = validatedValue.reason
+            }
+            is ValidationResult.ShouldTruncate -> {
+                valueToInsert = validatedValue.truncatedValue
+                changeReason = validatedValue.reason
+            }
+            ValidationResult.Valid -> {
+                valueToInsert = inputValue
+                changeReason = null
+            }
         }
+
+        client.createNamespace(testNamespace)
+        client.createTable(stream, tableName, columnNameMapping, replace = false)
+        testClient.insertRecords(
+            tableName,
+            columnNameMapping,
+            inputRecord("test" to valueToInsert),
+        )
+
+        val actualRecords =
+            testClient
+                .readTable(tableName)
+                .removeAirbyteColumns(airbyteMetaColumnMapping)
+                .reverseColumnNameMapping(columnNameMapping, airbyteMetaColumnMapping)
+                .removeNulls()
+        val actualValue = actualRecords.first()["test"]
+        assertEquals(
+            expectedValue,
+            actualValue,
+            "For input $inputValue, expected ${expectedValue.simpleClassName()}; actual value was ${actualValue.simpleClassName()}. Coercer output was $validatedValue.",
+        )
+        assertEquals(expectedChangeReason, changeReason)
     }
 }
+
+fun Any?.simpleClassName() = this?.let { it::class.simpleName } ?: "null"

airbyte-cdk/bulk/core/load/src/testFixtures/kotlin/io/airbyte/cdk/load/component/TableSchemaEvolutionSuite.kt

@@ -44,7 +44,13 @@ interface TableSchemaEvolutionSuite {
     val schemaFactory: TableSchemaFactory
 
     private val harness: TableOperationsTestHarness
-        get() = TableOperationsTestHarness(opsClient, testClient, airbyteMetaColumnMapping)
+        get() =
+            TableOperationsTestHarness(
+                opsClient,
+                testClient,
+                schemaFactory,
+                airbyteMetaColumnMapping
+            )
 
     /**
      * Test that the connector can correctly discover all of its own data types. This test creates a

airbyte-cdk/bulk/version.properties

@@ -1 +1 @@
-version=0.1.87
+version=0.1.91

airbyte-ci/connectors/auto_merge/src/auto_merge/consts.py

@@ -10,5 +10,6 @@ CONNECTOR_PATH_PREFIXES = {
     "airbyte-integrations/connectors",
     "docs/integrations/sources",
     "docs/integrations/destinations",
+    "docs/ai-agents/connectors",
 }
 MERGE_METHOD = "squash"

airbyte-ci/connectors/metadata_service/lib/README.md

@@ -75,7 +75,7 @@ This will copy the specified connector version to your development bucket. This
 _💡 Note: A prerequisite is you have [gsutil](https://cloud.google.com/storage/docs/gsutil) installed and have run `gsutil auth login`_
 
 ```bash
-TARGET_BUCKET=<YOUR-DEV_BUCKET> CONNECTOR="airbyte/source-stripe" VERSION="3.17.0-dev.ea013c8741" poetry run poe copy-connector-from-prod
+TARGET_BUCKET=<YOUR-DEV_BUCKET> CONNECTOR="airbyte/source-stripe" VERSION="3.17.0-preview.ea013c8" poetry run poe copy-connector-from-prod
 ```
 
 ### Promote Connector Version to Latest
@@ -87,5 +87,5 @@ _💡 Note: A prerequisite is you have [gsutil](https://cloud.google.com/storage
 _⚠️ Warning: Its important to know that this will remove ANY existing files in the latest folder that are not in the versioned folder as it calls `gsutil rsync` with `-d` enabled._
 
 ```bash
-TARGET_BUCKET=<YOUR-DEV_BUCKET> CONNECTOR="airbyte/source-stripe" VERSION="3.17.0-dev.ea013c8741" poetry run poe promote-connector-to-latest
+TARGET_BUCKET=<YOUR-DEV_BUCKET> CONNECTOR="airbyte/source-stripe" VERSION="3.17.0-preview.ea013c8" poetry run poe promote-connector-to-latest
 ```

airbyte-ci/connectors/metadata_service/lib/metadata_service/docker_hub.py

@@ -28,8 +28,8 @@ def get_docker_hub_auth_token() -> str:
 
 
 def get_docker_hub_headers() -> Dict | None:
-    if "DOCKER_HUB_USERNAME" not in os.environ or "DOCKER_HUB_PASSWORD" not in os.environ:
-        # If the Docker Hub credentials are not provided, we can only anonymously call the Docker Hub API.
+    if not os.environ.get("DOCKER_HUB_USERNAME") or not os.environ.get("DOCKER_HUB_PASSWORD"):
+        # If the Docker Hub credentials are not provided (or are empty), we can only anonymously call the Docker Hub API.
         # This will only work for public images and lead to a lower rate limit.
         return {}
     else:

airbyte-ci/connectors/metadata_service/lib/metadata_service/registry_entry.py

@@ -434,7 +434,7 @@ def generate_and_persist_registry_entry(
         bucket_name (str): The name of the GCS bucket.
         repo_metadata_file_path (pathlib.Path): The path to the spec file.
         registry_type (str): The registry type.
-        docker_image_tag (str): The docker image tag associated with this release. Typically a semver string (e.g. '1.2.3'), possibly with a suffix (e.g. '1.2.3-dev.abcde12345')
+        docker_image_tag (str): The docker image tag associated with this release. Typically a semver string (e.g. '1.2.3'), possibly with a suffix (e.g. '1.2.3-preview.abcde12')
         is_prerelease (bool): Whether this is a prerelease, or a main release.
     """
     # Read the repo metadata dict to bootstrap ourselves. We need the docker repository,
@@ -444,7 +444,7 @@ def generate_and_persist_registry_entry(
 
     try:
         # Now that we have the docker repo, read the appropriate versioned metadata from GCS.
-        # This metadata will differ in a few fields (e.g. in prerelease mode, dockerImageTag will contain the actual prerelease tag `1.2.3-dev.abcde12345`),
+        # This metadata will differ in a few fields (e.g. in prerelease mode, dockerImageTag will contain the actual prerelease tag `1.2.3-preview.abcde12`),
         # so we'll treat this as the source of truth (ish. See below for how we handle the registryOverrides field.)
         gcs_client = get_gcs_storage_client(gcs_creds=os.environ.get("GCS_CREDENTIALS"))
         bucket = gcs_client.bucket(bucket_name)
@@ -533,7 +533,9 @@ def generate_and_persist_registry_entry(
 
     # For latest versions that are disabled, delete any existing registry entry to remove it from the registry
     if (
-        "-rc" not in metadata_dict["data"]["dockerImageTag"] and "-dev" not in metadata_dict["data"]["dockerImageTag"]
+        "-rc" not in metadata_dict["data"]["dockerImageTag"]
+        and "-dev" not in metadata_dict["data"]["dockerImageTag"]
+        and "-preview" not in metadata_dict["data"]["dockerImageTag"]
     ) and not metadata_dict["data"]["registryOverrides"][registry_type]["enabled"]:
         logger.info(
             f"{registry_type} is not enabled: deleting existing {registry_type} registry entry for {metadata_dict['data']['dockerRepository']} at latest path."

airbyte-ci/connectors/metadata_service/lib/tests/fixtures/metadata_validate/valid/with_optional_field/with_releases/metadata_breaking_change_prerelease.yaml

@@ -5,7 +5,7 @@ data:
   connectorType: source
   dockerRepository: airbyte/image-exists-1
   githubIssueLabel: source-alloydb-strict-encrypt
-  dockerImageTag: 2.0.0-dev.cf3628ccf3
+  dockerImageTag: 2.0.0-preview.cf3628c
   documentationUrl: https://docs.airbyte.com/integrations/sources/existingsource
   connectorSubtype: database
   releaseStage: generally_available

airbyte-ci/connectors/metadata_service/lib/tests/test_commands.py

@@ -231,7 +231,7 @@ def test_upload_prerelease(mocker, valid_metadata_yaml_files, tmp_path):
     mocker.patch.object(commands.click, "secho")
     mocker.patch.object(commands, "upload_metadata_to_gcs")
 
-    prerelease_tag = "0.3.0-dev.6d33165120"
+    prerelease_tag = "0.3.0-preview.6d33165"
     bucket = "my-bucket"
     metadata_file_path = valid_metadata_yaml_files[0]
     validator_opts = ValidatorOptions(docs_path=str(tmp_path), prerelease_tag=prerelease_tag)

airbyte-ci/connectors/metadata_service/lib/tests/test_gcs_upload.py

@@ -582,7 +582,7 @@ def test_upload_metadata_to_gcs_invalid_docker_images(mocker, invalid_metadata_u
 def test_upload_metadata_to_gcs_with_prerelease(mocker, valid_metadata_upload_files, tmp_path):
     mocker.spy(gcs_upload, "_file_upload")
     mocker.spy(gcs_upload, "upload_file_if_changed")
-    prerelease_image_tag = "1.5.6-dev.f80318f754"
+    prerelease_image_tag = "1.5.6-preview.f80318f"
 
     for valid_metadata_upload_file in valid_metadata_upload_files:
         tmp_metadata_file_path = tmp_path / "metadata.yaml"
@@ -701,7 +701,7 @@ def test_upload_metadata_to_gcs_release_candidate(mocker, get_fixture_path, tmp_
     )
     assert metadata.data.releases.rolloutConfiguration.enableProgressiveRollout
 
-    prerelease_tag = "1.5.6-dev.f80318f754" if prerelease else None
+    prerelease_tag = "1.5.6-preview.f80318f" if prerelease else None
 
     upload_info = gcs_upload.upload_metadata_to_gcs(
         "my_bucket",

airbyte-ci/connectors/pipelines/pipelines/airbyte_ci/connectors/publish/context.py

@@ -110,14 +110,14 @@ class PublishConnectorContext(ConnectorContext):
 
     @property
     def pre_release_suffix(self) -> str:
-        return self.git_revision[:10]
+        return self.git_revision[:7]
 
     @property
     def docker_image_tag(self) -> str:
         # get the docker image tag from the parent class
         metadata_tag = super().docker_image_tag
         if self.pre_release:
-            return f"{metadata_tag}-dev.{self.pre_release_suffix}"
+            return f"{metadata_tag}-preview.{self.pre_release_suffix}"
         else:
             return metadata_tag
 

airbyte-ci/connectors/pipelines/pipelines/airbyte_ci/connectors/test/steps/python_connectors.py

@@ -25,7 +25,7 @@ from pipelines.helpers.utils import raise_if_not_user
 from pipelines.models.steps import STEP_PARAMS, Step, StepResult
 
 # Pin the PyAirbyte version to avoid updates from breaking CI
-PYAIRBYTE_VERSION = "0.20.2"
+PYAIRBYTE_VERSION = "0.35.1"
 
 
 class PytestStep(Step, ABC):

airbyte-ci/connectors/pipelines/tests/test_tests/test_python_connectors.py

@@ -156,7 +156,8 @@ class TestPyAirbyteValidationTests:
         result = await PyAirbyteValidation(context_for_valid_connector)._run(mocker.MagicMock())
         assert isinstance(result, StepResult)
         assert result.status == StepStatus.SUCCESS
-        assert "Getting `spec` output from connector..." in result.stdout
+        # Verify the connector name appears in output (stable across PyAirbyte versions)
+        assert context_for_valid_connector.connector.technical_name in (result.stdout + result.stderr)
 
     async def test__run_validation_skip_unpublished_connector(
         self,

airbyte-integrations/connectors/destination-clickhouse/gradle.properties

@@ -1,2 +1,2 @@
-cdkVersion=0.1.86
+cdkVersion=0.1.89
 JunitMethodExecutionTimeout=10m

airbyte-integrations/connectors/destination-clickhouse/metadata.yaml

@@ -2,7 +2,7 @@ data:
   connectorSubtype: database
   connectorType: destination
   definitionId: ce0d828e-1dc4-496c-b122-2da42e637e48
-  dockerImageTag: 2.1.16-rc.2
+  dockerImageTag: 2.1.18
   dockerRepository: airbyte/destination-clickhouse
   githubIssueLabel: destination-clickhouse
   icon: clickhouse.svg
@@ -27,7 +27,7 @@ data:
   releaseStage: generally_available
   releases:
     rolloutConfiguration:
-      enableProgressiveRollout: true
+      enableProgressiveRollout: false
     breakingChanges:
       2.0.0:
         message: "This connector has been re-written from scratch. Data will now be typed and stored in final (non-raw) tables. The connector may require changes to its configuration to function properly and downstream pipelines may be affected. Warning: SSH tunneling is in Beta."

airbyte-integrations/connectors/destination-clickhouse/src/main/kotlin/io/airbyte/integrations/destination/clickhouse/client/ClickhouseSqlGenerator.kt

@@ -54,8 +54,11 @@ class ClickhouseSqlGenerator {
                     // Check if cursor column type is valid for ClickHouse ReplacingMergeTree
                     val cursor = tableSchema.getCursor().firstOrNull()
                     val cursorType = cursor?.let { finalSchema[it]?.type }
+
+                    val useCursorAsVersion =
+                        cursorType != null && isValidVersionColumn(cursor, cursorType)
                     val versionColumn =
-                        if (cursorType?.isValidVersionColumnType() ?: false) {
+                        if (useCursorAsVersion) {
                             "`$cursor`"
                         } else {
                             // Fallback to _airbyte_extracted_at if no cursor is specified or cursor

airbyte-integrations/connectors/destination-clickhouse/src/main/kotlin/io/airbyte/integrations/destination/clickhouse/client/ClickhouseSqlTypes.kt

@@ -4,6 +4,7 @@
 
 package io.airbyte.integrations.destination.clickhouse.client
 
+import io.airbyte.cdk.load.table.CDC_CURSOR_COLUMN
 import io.airbyte.integrations.destination.clickhouse.client.ClickhouseSqlTypes.VALID_VERSION_COLUMN_TYPES
 
 object ClickhouseSqlTypes {
@@ -23,4 +24,9 @@ object ClickhouseSqlTypes {
         )
 }
 
-fun String.isValidVersionColumnType() = VALID_VERSION_COLUMN_TYPES.contains(this)
+// Warning: if any munging changes the name of the CDC column name this will break.
+// Currently, that is not the case.
+fun isValidVersionColumn(name: String, type: String) =
+    // CDC cursors cannot be used as a version column since they are null
+    // during the initial CDC snapshot.
+    name != CDC_CURSOR_COLUMN && VALID_VERSION_COLUMN_TYPES.contains(type)

airbyte-integrations/connectors/destination-clickhouse/src/main/kotlin/io/airbyte/integrations/destination/clickhouse/config/ClickhouseNameGenerators.kt

@@ -1,62 +0,0 @@
-/*
- * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
- */
-
-package io.airbyte.integrations.destination.clickhouse.config
-
-import io.airbyte.cdk.load.command.DestinationStream
-import io.airbyte.cdk.load.data.Transformations.Companion.toAlphanumericAndUnderscore
-import io.airbyte.cdk.load.schema.model.TableName
-import io.airbyte.cdk.load.table.ColumnNameGenerator
-import io.airbyte.cdk.load.table.FinalTableNameGenerator
-import io.airbyte.integrations.destination.clickhouse.spec.ClickhouseConfiguration
-import jakarta.inject.Singleton
-import java.util.Locale
-import java.util.UUID
-
-@Singleton
-class ClickhouseFinalTableNameGenerator(private val config: ClickhouseConfiguration) :
-    FinalTableNameGenerator {
-    override fun getTableName(streamDescriptor: DestinationStream.Descriptor) =
-        TableName(
-            namespace =
-                (streamDescriptor.namespace ?: config.resolvedDatabase)
-                    .toClickHouseCompatibleName(),
-            name = streamDescriptor.name.toClickHouseCompatibleName(),
-        )
-}
-
-@Singleton
-class ClickhouseColumnNameGenerator : ColumnNameGenerator {
-    override fun getColumnName(column: String): ColumnNameGenerator.ColumnName {
-        return ColumnNameGenerator.ColumnName(
-            column.toClickHouseCompatibleName(),
-            column.lowercase(Locale.getDefault()).toClickHouseCompatibleName(),
-        )
-    }
-}
-
-/**
- * Transforms a string to be compatible with ClickHouse table and column names.
- *
- * @return The transformed string suitable for ClickHouse identifiers.
- */
-fun String.toClickHouseCompatibleName(): String {
-    // 1. Replace any character that is not a letter,
-    //    a digit (0-9), or an underscore (_) with a single underscore.
-    var transformed = toAlphanumericAndUnderscore(this)
-
-    // 2. Ensure the identifier does not start with a digit.
-    //    If it starts with a digit, prepend an underscore.
-    if (transformed.isNotEmpty() && transformed[0].isDigit()) {
-        transformed = "_$transformed"
-    }
-
-    // 3.Do not allow empty strings.
-    if (transformed.isEmpty()) {
-        return "default_name_${UUID.randomUUID()}" // A fallback name if the input results in an
-        // empty string
-    }
-
-    return transformed
-}

airbyte-integrations/connectors/destination-clickhouse/src/main/kotlin/io/airbyte/integrations/destination/clickhouse/schema/ClickhouseNamingUtils.kt

@@ -0,0 +1,33 @@
+/*
+ * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
+ */
+
+package io.airbyte.integrations.destination.clickhouse.schema
+
+import io.airbyte.cdk.load.data.Transformations.Companion.toAlphanumericAndUnderscore
+import java.util.UUID
+
+/**
+ * Transforms a string to be compatible with ClickHouse table and column names.
+ *
+ * @return The transformed string suitable for ClickHouse identifiers.
+ */
+fun String.toClickHouseCompatibleName(): String {
+    // 1. Replace any character that is not a letter,
+    //    a digit (0-9), or an underscore (_) with a single underscore.
+    var transformed = toAlphanumericAndUnderscore(this)
+
+    // 2.Do not allow empty strings.
+    if (transformed.isEmpty()) {
+        return "default_name_${UUID.randomUUID()}" // A fallback name if the input results in an
+        // empty string
+    }
+
+    // 3. Ensure the identifier does not start with a digit.
+    //    If it starts with a digit, prepend an underscore.
+    if (transformed[0].isDigit()) {
+        transformed = "_$transformed"
+    }
+
+    return transformed
+}

airbyte-integrations/connectors/destination-clickhouse/src/main/kotlin/io/airbyte/integrations/destination/clickhouse/schema/ClickhouseTableSchemaMapper.kt

@@ -29,8 +29,7 @@ import io.airbyte.cdk.load.schema.model.StreamTableSchema
 import io.airbyte.cdk.load.schema.model.TableName
 import io.airbyte.cdk.load.table.TempTableNameGenerator
 import io.airbyte.integrations.destination.clickhouse.client.ClickhouseSqlTypes
-import io.airbyte.integrations.destination.clickhouse.client.isValidVersionColumnType
-import io.airbyte.integrations.destination.clickhouse.config.toClickHouseCompatibleName
+import io.airbyte.integrations.destination.clickhouse.client.isValidVersionColumn
 import io.airbyte.integrations.destination.clickhouse.spec.ClickhouseConfiguration
 import jakarta.inject.Singleton
 
@@ -100,7 +99,7 @@ class ClickhouseTableSchemaMapper(
             if (cursor != null) {
                 // Check if the cursor column type is valid for ClickHouse ReplacingMergeTree
                 val cursorColumnType = tableSchema.columnSchema.finalSchema[cursor]!!.type
-                if (cursorColumnType.isValidVersionColumnType()) {
+                if (isValidVersionColumn(cursor, cursorColumnType)) {
                     // Cursor column is valid, use it as version column
                     add(cursor) // Make cursor column non-nullable too
                 }

airbyte-integrations/connectors/destination-clickhouse/src/test-integration/kotlin/io/airbyte/integrations/destination/clickhouse/component/ClickhouseDataCoercionTest.kt

@@ -0,0 +1,77 @@
+/*
+ * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
+ */
+
+package io.airbyte.integrations.destination.clickhouse.component
+
+import io.airbyte.cdk.load.component.DataCoercionNumberFixtures
+import io.airbyte.cdk.load.component.DataCoercionNumberFixtures.NEGATIVE_HIGH_PRECISION_FLOAT
+import io.airbyte.cdk.load.component.DataCoercionNumberFixtures.POSITIVE_HIGH_PRECISION_FLOAT
+import io.airbyte.cdk.load.component.DataCoercionNumberFixtures.SMALLEST_NEGATIVE_FLOAT32
+import io.airbyte.cdk.load.component.DataCoercionNumberFixtures.SMALLEST_NEGATIVE_FLOAT64
+import io.airbyte.cdk.load.component.DataCoercionNumberFixtures.SMALLEST_POSITIVE_FLOAT32
+import io.airbyte.cdk.load.component.DataCoercionNumberFixtures.SMALLEST_POSITIVE_FLOAT64
+import io.airbyte.cdk.load.component.DataCoercionSuite
+import io.airbyte.cdk.load.component.TableOperationsClient
+import io.airbyte.cdk.load.component.TestTableOperationsClient
+import io.airbyte.cdk.load.component.toArgs
+import io.airbyte.cdk.load.data.AirbyteValue
+import io.airbyte.cdk.load.dataflow.transform.ValueCoercer
+import io.airbyte.cdk.load.schema.TableSchemaFactory
+import io.airbyte.protocol.models.v0.AirbyteRecordMessageMetaChange.Reason
+import io.micronaut.test.extensions.junit5.annotation.MicronautTest
+import org.junit.jupiter.params.ParameterizedTest
+import org.junit.jupiter.params.provider.MethodSource
+
+@MicronautTest(environments = ["component"], resolveParameters = false)
+class ClickhouseDataCoercionTest(
+    override val coercer: ValueCoercer,
+    override val opsClient: TableOperationsClient,
+    override val testClient: TestTableOperationsClient,
+    override val schemaFactory: TableSchemaFactory,
+) : DataCoercionSuite {
+    @ParameterizedTest
+    // We use clickhouse's Int64 type for integers
+    @MethodSource("io.airbyte.cdk.load.component.DataCoercionIntegerFixtures#int64")
+    override fun `handle integer values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) {
+        super.`handle integer values`(inputValue, expectedValue, expectedChangeReason)
+    }
+
+    @ParameterizedTest
+    @MethodSource(
+        "io.airbyte.integrations.destination.clickhouse.component.ClickhouseDataCoercionTest#numbers"
+    )
+    override fun `handle number values`(
+        inputValue: AirbyteValue,
+        expectedValue: Any?,
+        expectedChangeReason: Reason?
+    ) {
+        super.`handle number values`(inputValue, expectedValue, expectedChangeReason)
+    }
+
+    companion object {
+        /**
+         * destination-clickhouse doesn't set a change reason when truncating high-precision numbers
+         * (https://github.com/airbytehq/airbyte-internal-issues/issues/15401)
+         */
+        @JvmStatic
+        fun numbers() =
+            DataCoercionNumberFixtures.numeric38_9
+                .map {
+                    when (it.name) {
+                        POSITIVE_HIGH_PRECISION_FLOAT,
+                        NEGATIVE_HIGH_PRECISION_FLOAT,
+                        SMALLEST_POSITIVE_FLOAT32,
+                        SMALLEST_NEGATIVE_FLOAT32,
+                        SMALLEST_POSITIVE_FLOAT64,
+                        SMALLEST_NEGATIVE_FLOAT64 -> it.copy(changeReason = null)
+                        else -> it
+                    }
+                }
+                .toArgs()
+    }
+}

airbyte-integrations/connectors/destination-clickhouse/src/test-integration/kotlin/io/airbyte/integrations/destination/clickhouse/component/ClickhouseTableSchemaEvolutionTest.kt

@@ -22,7 +22,7 @@ class ClickhouseTableSchemaEvolutionTest(
     override val client: TableSchemaEvolutionClient,
     override val opsClient: TableOperationsClient,
     override val testClient: TestTableOperationsClient,
-    override val schemaFactory: TableSchemaFactory
+    override val schemaFactory: TableSchemaFactory,
 ) : TableSchemaEvolutionSuite {
     private val allTypesTableSchema =
         TableSchema(

airbyte-integrations/connectors/destination-clickhouse/src/test-integration/kotlin/io/airbyte/integrations/destination/clickhouse/fixtures/ClickhouseExpectedRecordMapper.kt

@@ -16,7 +16,7 @@ import io.airbyte.cdk.load.data.TimestampWithTimezoneValue
 import io.airbyte.cdk.load.data.TimestampWithoutTimezoneValue
 import io.airbyte.cdk.load.test.util.ExpectedRecordMapper
 import io.airbyte.cdk.load.test.util.OutputRecord
-import io.airbyte.integrations.destination.clickhouse.config.toClickHouseCompatibleName
+import io.airbyte.integrations.destination.clickhouse.schema.toClickHouseCompatibleName
 import java.math.RoundingMode
 import java.time.LocalTime
 import java.time.ZoneOffset

airbyte-integrations/connectors/destination-clickhouse/src/test-integration/kotlin/io/airbyte/integrations/destination/clickhouse/write/load/ClickhouseAcceptanceTest.kt

@@ -30,8 +30,8 @@ import io.airbyte.cdk.load.write.UnknownTypesBehavior
 import io.airbyte.integrations.destination.clickhouse.ClickhouseConfigUpdater
 import io.airbyte.integrations.destination.clickhouse.ClickhouseContainerHelper
 import io.airbyte.integrations.destination.clickhouse.Utils
-import io.airbyte.integrations.destination.clickhouse.config.toClickHouseCompatibleName
 import io.airbyte.integrations.destination.clickhouse.fixtures.ClickhouseExpectedRecordMapper
+import io.airbyte.integrations.destination.clickhouse.schema.toClickHouseCompatibleName
 import io.airbyte.integrations.destination.clickhouse.spec.ClickhouseConfiguration
 import io.airbyte.integrations.destination.clickhouse.spec.ClickhouseConfigurationFactory
 import io.airbyte.integrations.destination.clickhouse.spec.ClickhouseSpecificationOss

airbyte-integrations/connectors/destination-clickhouse/src/test/kotlin/io/airbyte/integrations/destination/clickhouse/client/ClickhouseAirbyteClientTest.kt

@@ -21,7 +21,6 @@ import io.airbyte.cdk.load.schema.model.StreamTableSchema
 import io.airbyte.cdk.load.schema.model.TableName
 import io.airbyte.cdk.load.table.ColumnNameMapping
 import io.airbyte.cdk.load.table.TempTableNameGenerator
-import io.airbyte.integrations.destination.clickhouse.config.ClickhouseFinalTableNameGenerator
 import io.mockk.coEvery
 import io.mockk.coVerify
 import io.mockk.coVerifyOrder
@@ -39,8 +38,6 @@ class ClickhouseAirbyteClientTest {
     // Mocks
     private val client: ClickHouseClientRaw = mockk(relaxed = true)
     private val clickhouseSqlGenerator: ClickhouseSqlGenerator = mockk(relaxed = true)
-    private val clickhouseFinalTableNameGenerator: ClickhouseFinalTableNameGenerator =
-        mockk(relaxed = true)
     private val tempTableNameGenerator: TempTableNameGenerator = mockk(relaxed = true)
 
     // Client
@@ -105,7 +102,6 @@ class ClickhouseAirbyteClientTest {
             alterTableStatement
         coEvery { clickhouseAirbyteClient.execute(alterTableStatement) } returns
             mockk(relaxed = true)
-        every { clickhouseFinalTableNameGenerator.getTableName(any()) } returns mockTableName
 
         mockCHSchemaWithAirbyteColumns()
 
@@ -172,7 +168,6 @@ class ClickhouseAirbyteClientTest {
 
         coEvery { clickhouseAirbyteClient.execute(any()) } returns mockk(relaxed = true)
         every { tempTableNameGenerator.generate(any()) } returns tempTableName
-        every { clickhouseFinalTableNameGenerator.getTableName(any()) } returns finalTableName
 
         mockCHSchemaWithAirbyteColumns()
 
@@ -226,8 +221,6 @@ class ClickhouseAirbyteClientTest {
     fun `test ensure schema matches fails if no airbyte columns`() = runTest {
         val finalTableName = TableName("fin", "al")
 
-        every { clickhouseFinalTableNameGenerator.getTableName(any()) } returns finalTableName
-
         val columnMapping = ColumnNameMapping(mapOf())
         val stream =
             mockk<DestinationStream> {

airbyte-integrations/connectors/destination-clickhouse/src/test/kotlin/io/airbyte/integrations/destination/clickhouse/schema/ClickhouseNamingUtilsTest.kt

@@ -2,13 +2,13 @@
  * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
  */
 
-package io.airbyte.integrations.destination.clickhouse.config
+package io.airbyte.integrations.destination.clickhouse.schema
 
 import java.util.UUID
 import org.junit.jupiter.api.Assertions
 import org.junit.jupiter.api.Test
 
-class ClickhouseNameGeneratorTest {
+class ClickhouseNamingUtilsTest {
     @Test
     fun `toClickHouseCompatibleName replaces special characters with underscores`() {
         Assertions.assertEquals("hello_world", "hello world".toClickHouseCompatibleName())

airbyte-integrations/connectors/destination-postgres/metadata.yaml

@@ -6,7 +6,7 @@ data:
   connectorSubtype: database
   connectorType: destination
   definitionId: 25c5221d-dce2-4163-ade9-739ef790f503
-  dockerImageTag: 3.0.5-rc.1
+  dockerImageTag: 3.0.5
   dockerRepository: airbyte/destination-postgres
   documentationUrl: https://docs.airbyte.com/integrations/destinations/postgres
   githubIssueLabel: destination-postgres
@@ -22,7 +22,7 @@ data:
       enabled: true
   releases:
     rolloutConfiguration:
-      enableProgressiveRollout: true
+      enableProgressiveRollout: false
     breakingChanges:
       3.0.0:
         message: >

airbyte-integrations/connectors/destination-postgres/src/main/kotlin/io/airbyte/integrations/destination/postgres/client/PostgresAirbyteClient.kt

@@ -4,12 +4,16 @@
 
 package io.airbyte.integrations.destination.postgres.client
 
+import edu.umd.cs.findbugs.annotations.SuppressFBWarnings
+import io.airbyte.cdk.ConfigErrorException
 import io.airbyte.cdk.load.command.DestinationStream
 import io.airbyte.cdk.load.component.ColumnChangeset
+import io.airbyte.cdk.load.component.ColumnType
 import io.airbyte.cdk.load.component.TableColumns
 import io.airbyte.cdk.load.component.TableOperationsClient
 import io.airbyte.cdk.load.component.TableSchema
 import io.airbyte.cdk.load.component.TableSchemaEvolutionClient
+import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAMES
 import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_GENERATION_ID
 import io.airbyte.cdk.load.schema.model.TableName
 import io.airbyte.cdk.load.table.ColumnNameMapping
@@ -26,6 +30,11 @@ import javax.sql.DataSource
 private val log = KotlinLogging.logger {}
 
 @Singleton
+@SuppressFBWarnings(
+    value = ["SQL_NONCONSTANT_STRING_PASSED_TO_EXECUTE"],
+    justification =
+        "There is little chance of SQL injection. There is also little need for statement reuse. The basic statement is more readable than the prepared statement."
+)
 class PostgresAirbyteClient(
     private val dataSource: DataSource,
     private val sqlGenerator: PostgresDirectLoadSqlGenerator,
@@ -53,6 +62,29 @@ class PostgresAirbyteClient(
             null
         }
 
+    override suspend fun namespaceExists(namespace: String): Boolean {
+        return executeQuery(
+            """
+            SELECT EXISTS(
+                SELECT 1 FROM information_schema.schemata
+                WHERE schema_name = '$namespace'
+            )
+            """
+        ) { rs -> rs.next() && rs.getBoolean(1) }
+    }
+
+    override suspend fun tableExists(table: TableName): Boolean {
+        return executeQuery(
+            """
+            SELECT EXISTS(
+                SELECT 1 FROM information_schema.tables
+                WHERE table_schema = '${table.namespace}'
+                AND table_name = '${table.name}'
+            )
+            """
+        ) { rs -> rs.next() && rs.getBoolean(1) }
+    }
+
     override suspend fun createNamespace(namespace: String) {
         try {
             execute(sqlGenerator.createNamespace(namespace))
@@ -171,14 +203,26 @@ class PostgresAirbyteClient(
     }
 
     override suspend fun discoverSchema(tableName: TableName): TableSchema {
-        TODO("Not yet implemented")
+        val columnsInDb = getColumnsFromDbForDiscovery(tableName)
+        val hasAllAirbyteColumns = columnsInDb.keys.containsAll(COLUMN_NAMES)
+
+        if (!hasAllAirbyteColumns) {
+            val message =
+                "The target table ($tableName) already exists in the destination, but does not contain Airbyte's internal columns. Airbyte can only sync to Airbyte-controlled tables. To fix this error, you must either delete the target table or add a prefix in the connection configuration in order to sync to a separate table in the destination."
+            log.error { message }
+            throw ConfigErrorException(message)
+        }
+
+        // Filter out Airbyte columns
+        val userColumns = columnsInDb.filterKeys { it !in COLUMN_NAMES }
+        return TableSchema(userColumns)
     }
 
     override fun computeSchema(
         stream: DestinationStream,
         columnNameMapping: ColumnNameMapping
     ): TableSchema {
-        TODO("Not yet implemented")
+        return TableSchema(stream.tableSchema.columnSchema.finalSchema)
     }
 
     override suspend fun applyChangeset(
@@ -188,9 +232,73 @@ class PostgresAirbyteClient(
         expectedColumns: TableColumns,
         columnChangeset: ColumnChangeset
     ) {
-        TODO("Not yet implemented")
+        if (
+            columnChangeset.columnsToAdd.isNotEmpty() ||
+                columnChangeset.columnsToDrop.isNotEmpty() ||
+                columnChangeset.columnsToChange.isNotEmpty()
+        ) {
+            log.info { "Summary of the table alterations:" }
+            log.info { "Added columns: ${columnChangeset.columnsToAdd}" }
+            log.info { "Deleted columns: ${columnChangeset.columnsToDrop}" }
+            log.info { "Modified columns: ${columnChangeset.columnsToChange}" }
+
+            // Convert from TableColumns format to Column format
+            val columnsToAdd =
+                columnChangeset.columnsToAdd
+                    .map { (name, type) -> Column(name, type.type, type.nullable) }
+                    .toSet()
+            val columnsToRemove =
+                columnChangeset.columnsToDrop
+                    .map { (name, type) -> Column(name, type.type, type.nullable) }
+                    .toSet()
+            val columnsToModify =
+                columnChangeset.columnsToChange
+                    .map { (name, change) ->
+                        Column(name, change.newType.type, change.newType.nullable)
+                    }
+                    .toSet()
+            val columnsInDb =
+                (columnChangeset.columnsToRetain +
+                        columnChangeset.columnsToDrop +
+                        columnChangeset.columnsToChange.mapValues { it.value.originalType })
+                    .map { (name, type) -> Column(name, type.type, type.nullable) }
+                    .toSet()
+
+            execute(
+                sqlGenerator.matchSchemas(
+                    tableName = tableName,
+                    columnsToAdd = columnsToAdd,
+                    columnsToRemove = columnsToRemove,
+                    columnsToModify = columnsToModify,
+                    columnsInDb = columnsInDb,
+                    recreatePrimaryKeyIndex = false,
+                    primaryKeyColumnNames = emptyList(),
+                    recreateCursorIndex = false,
+                    cursorColumnName = null,
+                )
+            )
+        }
     }
 
+    /**
+     * Gets columns from the database including their types for schema discovery. Unlike
+     * [getColumnsFromDb], this returns all columns including Airbyte metadata columns.
+     */
+    private fun getColumnsFromDbForDiscovery(tableName: TableName): Map<String, ColumnType> =
+        executeQuery(sqlGenerator.getTableSchema(tableName)) { rs ->
+            val columnsInDb: MutableMap<String, ColumnType> = mutableMapOf()
+            while (rs.next()) {
+                val columnName = rs.getString(COLUMN_NAME_COLUMN)
+                val dataType = rs.getString("data_type")
+                // PostgreSQL's information_schema always returns 'YES' or 'NO' for is_nullable
+                val isNullable = rs.getString("is_nullable") == "YES"
+
+                columnsInDb[columnName] = ColumnType(normalizePostgresType(dataType), isNullable)
+            }
+
+            columnsInDb
+        }
+
     /**
      * Checks if the primary key index matches the current stream configuration. If the primary keys
      * have changed (detected by comparing columns in the index), then this will return true,

airbyte-integrations/connectors/destination-postgres/src/main/kotlin/io/airbyte/integrations/destination/postgres/sql/PostgresDirectLoadSqlGenerator.kt

@@ -531,7 +531,7 @@ class PostgresDirectLoadSqlGenerator(
 
     fun getTableSchema(tableName: TableName): String =
         """
-        SELECT column_name, data_type
+        SELECT column_name, data_type, is_nullable
         FROM information_schema.columns
         WHERE table_schema = '${tableName.namespace}'
         AND table_name = '${tableName.name}';

airbyte-integrations/connectors/destination-postgres/src/main/kotlin/io/airbyte/integrations/destination/postgres/write/PostgresWriter.kt

@@ -49,6 +49,7 @@ class PostgresWriter(
     override fun createStreamLoader(stream: DestinationStream): StreamLoader {
         val initialStatus = initialStatuses[stream]!!
         val realTableName = stream.tableSchema.tableNames.finalTableName!!
+
         val tempTableName = tempTableNameGenerator.generate(realTableName)
         val columnNameMapping =
             ColumnNameMapping(stream.tableSchema.columnSchema.inputToFinalColumnNames)

airbyte-integrations/connectors/destination-postgres/src/test-integration/kotlin/io/airbyte/integrations/destination/postgres/component/PostgresComponentTestConfigFactory.kt

@@ -0,0 +1,45 @@
+/*
+ * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
+ */
+
+package io.airbyte.integrations.destination.postgres.component
+
+import io.airbyte.cdk.load.util.Jsons
+import io.airbyte.integrations.destination.postgres.PostgresConfigUpdater
+import io.airbyte.integrations.destination.postgres.PostgresContainerHelper
+import io.airbyte.integrations.destination.postgres.spec.PostgresConfiguration
+import io.airbyte.integrations.destination.postgres.spec.PostgresConfigurationFactory
+import io.airbyte.integrations.destination.postgres.spec.PostgresSpecificationOss
+import io.micronaut.context.annotation.Factory
+import io.micronaut.context.annotation.Primary
+import io.micronaut.context.annotation.Requires
+import jakarta.inject.Singleton
+
+@Requires(env = ["component"])
+@Factory
+class PostgresComponentTestConfigFactory {
+    @Singleton
+    @Primary
+    fun config(): PostgresConfiguration {
+        // Start the postgres container
+        PostgresContainerHelper.start()
+
+        // Create a minimal config JSON and update it with container details
+        val configJson =
+            """
+            {
+                "host": "replace_me_host",
+                "port": "replace_me_port",
+                "database": "replace_me_database",
+                "schema": "public",
+                "username": "replace_me_username",
+                "password": "replace_me_password",
+                "ssl": false
+            }
+        """
+
+        val updatedConfig = PostgresConfigUpdater().update(configJson)
+        val spec = Jsons.readValue(updatedConfig, PostgresSpecificationOss::class.java)
+        return PostgresConfigurationFactory().makeWithoutExceptionHandling(spec)
+    }
+}

airbyte-integrations/connectors/destination-postgres/src/test-integration/kotlin/io/airbyte/integrations/destination/postgres/component/PostgresComponentTestFixtures.kt

@@ -0,0 +1,36 @@
+/*
+ * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
+ */
+
+package io.airbyte.integrations.destination.postgres.component
+
+import io.airbyte.cdk.load.component.ColumnType
+import io.airbyte.cdk.load.component.TableOperationsFixtures
+import io.airbyte.cdk.load.component.TableSchema
+
+object PostgresComponentTestFixtures {
+    // PostgreSQL uses lowercase column names by default (no transformation needed)
+    val testMapping = TableOperationsFixtures.TEST_MAPPING
+    val idAndTestMapping = TableOperationsFixtures.ID_AND_TEST_MAPPING
+    val idTestWithCdcMapping = TableOperationsFixtures.ID_TEST_WITH_CDC_MAPPING
+
+    val allTypesTableSchema =
+        TableSchema(
+            mapOf(
+                "string" to ColumnType("varchar", true),
+                "boolean" to ColumnType("boolean", true),
+                "integer" to ColumnType("bigint", true),
+                "number" to ColumnType("decimal", true),
+                "date" to ColumnType("date", true),
+                "timestamp_tz" to ColumnType("timestamp with time zone", true),
+                "timestamp_ntz" to ColumnType("timestamp", true),
+                "time_tz" to ColumnType("time with time zone", true),
+                "time_ntz" to ColumnType("time", true),
+                "array" to ColumnType("jsonb", true),
+                "object" to ColumnType("jsonb", true),
+                "unknown" to ColumnType("jsonb", true),
+            )
+        )
+
+    val allTypesColumnNameMapping = TableOperationsFixtures.ALL_TYPES_MAPPING
+}

airbyte-integrations/connectors/destination-postgres/src/test-integration/kotlin/io/airbyte/integrations/destination/postgres/component/PostgresTableOperationsTest.kt

@@ -0,0 +1,92 @@
+/*
+ * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
+ */
+
+package io.airbyte.integrations.destination.postgres.component
+
+import io.airbyte.cdk.load.component.TableOperationsFixtures
+import io.airbyte.cdk.load.component.TableOperationsSuite
+import io.airbyte.cdk.load.schema.TableSchemaFactory
+import io.airbyte.integrations.destination.postgres.client.PostgresAirbyteClient
+import io.airbyte.integrations.destination.postgres.component.PostgresComponentTestFixtures.idTestWithCdcMapping
+import io.airbyte.integrations.destination.postgres.component.PostgresComponentTestFixtures.testMapping
+import io.micronaut.test.extensions.junit5.annotation.MicronautTest
+import jakarta.inject.Inject
+import org.junit.jupiter.api.Disabled
+import org.junit.jupiter.api.Test
+
+@MicronautTest(environments = ["component"])
+class PostgresTableOperationsTest(
+    override val client: PostgresAirbyteClient,
+    override val testClient: PostgresTestTableOperationsClient,
+) : TableOperationsSuite {
+
+    @Inject override lateinit var schemaFactory: TableSchemaFactory
+
+    @Test
+    override fun `connect to database`() {
+        super.`connect to database`()
+    }
+
+    @Test
+    override fun `create and drop namespaces`() {
+        super.`create and drop namespaces`()
+    }
+
+    @Test
+    override fun `create and drop tables`() {
+        super.`create and drop tables`()
+    }
+
+    @Test
+    override fun `insert records`() {
+        super.`insert records`(
+            inputRecords = TableOperationsFixtures.SINGLE_TEST_RECORD_INPUT,
+            expectedRecords = TableOperationsFixtures.SINGLE_TEST_RECORD_EXPECTED,
+            columnNameMapping = testMapping,
+        )
+    }
+
+    @Test
+    override fun `count table rows`() {
+        super.`count table rows`(columnNameMapping = testMapping)
+    }
+
+    @Test
+    override fun `overwrite tables`() {
+        super.`overwrite tables`(
+            sourceInputRecords = TableOperationsFixtures.OVERWRITE_SOURCE_RECORDS,
+            targetInputRecords = TableOperationsFixtures.OVERWRITE_TARGET_RECORDS,
+            expectedRecords = TableOperationsFixtures.OVERWRITE_EXPECTED_RECORDS,
+            columnNameMapping = testMapping,
+        )
+    }
+
+    @Test
+    override fun `copy tables`() {
+        super.`copy tables`(
+            sourceInputRecords = TableOperationsFixtures.OVERWRITE_SOURCE_RECORDS,
+            targetInputRecords = TableOperationsFixtures.OVERWRITE_TARGET_RECORDS,
+            expectedRecords = TableOperationsFixtures.COPY_EXPECTED_RECORDS,
+            columnNameMapping = testMapping,
+        )
+    }
+
+    @Test
+    override fun `get generation id`() {
+        super.`get generation id`(columnNameMapping = testMapping)
+    }
+
+    // TODO: Re-enable when CDK TableOperationsSuite is fixed to use ID_AND_TEST_SCHEMA for target
+    // table instead of TEST_INTEGER_SCHEMA (the Dedupe mode requires the id column as primary key)
+    @Disabled("CDK TableOperationsSuite bug: target table schema missing 'id' column for Dedupe")
+    @Test
+    override fun `upsert tables`() {
+        super.`upsert tables`(
+            sourceInputRecords = TableOperationsFixtures.UPSERT_SOURCE_RECORDS,
+            targetInputRecords = TableOperationsFixtures.UPSERT_TARGET_RECORDS,
+            expectedRecords = TableOperationsFixtures.UPSERT_EXPECTED_RECORDS,
+            columnNameMapping = idTestWithCdcMapping,
+        )
+    }
+}

airbyte-integrations/connectors/destination-postgres/src/test-integration/kotlin/io/airbyte/integrations/destination/postgres/component/PostgresTableSchemaEvolutionTest.kt

@@ -0,0 +1,111 @@
+/*
+ * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
+ */
+
+package io.airbyte.integrations.destination.postgres.component
+
+import io.airbyte.cdk.load.command.ImportType
+import io.airbyte.cdk.load.component.TableSchemaEvolutionFixtures
+import io.airbyte.cdk.load.component.TableSchemaEvolutionSuite
+import io.airbyte.cdk.load.schema.TableSchemaFactory
+import io.airbyte.integrations.destination.postgres.client.PostgresAirbyteClient
+import io.airbyte.integrations.destination.postgres.component.PostgresComponentTestFixtures.allTypesColumnNameMapping
+import io.airbyte.integrations.destination.postgres.component.PostgresComponentTestFixtures.allTypesTableSchema
+import io.airbyte.integrations.destination.postgres.component.PostgresComponentTestFixtures.idAndTestMapping
+import io.airbyte.integrations.destination.postgres.component.PostgresComponentTestFixtures.testMapping
+import io.micronaut.test.extensions.junit5.annotation.MicronautTest
+import org.junit.jupiter.api.Test
+
+@MicronautTest(environments = ["component"], resolveParameters = false)
+class PostgresTableSchemaEvolutionTest(
+    override val client: PostgresAirbyteClient,
+    override val opsClient: PostgresAirbyteClient,
+    override val testClient: PostgresTestTableOperationsClient,
+    override val schemaFactory: TableSchemaFactory,
+) : TableSchemaEvolutionSuite {
+
+    @Test
+    fun `discover recognizes all data types`() {
+        super.`discover recognizes all data types`(allTypesTableSchema, allTypesColumnNameMapping)
+    }
+
+    @Test
+    fun `computeSchema handles all data types`() {
+        super.`computeSchema handles all data types`(allTypesTableSchema, allTypesColumnNameMapping)
+    }
+
+    @Test
+    override fun `noop diff`() {
+        super.`noop diff`(testMapping)
+    }
+
+    @Test
+    override fun `changeset is correct when adding a column`() {
+        super.`changeset is correct when adding a column`(testMapping, idAndTestMapping)
+    }
+
+    @Test
+    override fun `changeset is correct when dropping a column`() {
+        super.`changeset is correct when dropping a column`(idAndTestMapping, testMapping)
+    }
+
+    @Test
+    override fun `changeset is correct when changing a column's type`() {
+        super.`changeset is correct when changing a column's type`(testMapping)
+    }
+
+    @Test
+    override fun `apply changeset - handle sync mode append`() {
+        super.`apply changeset - handle sync mode append`()
+    }
+
+    @Test
+    override fun `apply changeset - handle changing sync mode from append to dedup`() {
+        super.`apply changeset - handle changing sync mode from append to dedup`()
+    }
+
+    @Test
+    override fun `apply changeset - handle changing sync mode from dedup to append`() {
+        super.`apply changeset - handle changing sync mode from dedup to append`()
+    }
+
+    @Test
+    override fun `apply changeset - handle sync mode dedup`() {
+        super.`apply changeset - handle sync mode dedup`()
+    }
+
+    override fun `apply changeset`(
+        initialStreamImportType: ImportType,
+        modifiedStreamImportType: ImportType,
+    ) {
+        super.`apply changeset`(
+            initialColumnNameMapping =
+                TableSchemaEvolutionFixtures.APPLY_CHANGESET_INITIAL_COLUMN_MAPPING,
+            modifiedColumnNameMapping =
+                TableSchemaEvolutionFixtures.APPLY_CHANGESET_MODIFIED_COLUMN_MAPPING,
+            TableSchemaEvolutionFixtures.APPLY_CHANGESET_EXPECTED_EXTRACTED_AT,
+            initialStreamImportType,
+            modifiedStreamImportType,
+        )
+    }
+
+    @Test
+    override fun `change from string type to unknown type`() {
+        super.`change from string type to unknown type`(
+            idAndTestMapping,
+            idAndTestMapping,
+            TableSchemaEvolutionFixtures.STRING_TO_UNKNOWN_TYPE_INPUT_RECORDS,
+            TableSchemaEvolutionFixtures.STRING_TO_UNKNOWN_TYPE_EXPECTED_RECORDS,
+        )
+    }
+
+    @Test
+    override fun `change from unknown type to string type`() {
+        super.`change from unknown type to string type`(
+            idAndTestMapping,
+            idAndTestMapping,
+            TableSchemaEvolutionFixtures.UNKNOWN_TO_STRING_TYPE_INPUT_RECORDS,
+            TableSchemaEvolutionFixtures.UNKNOWN_TO_STRING_TYPE_EXPECTED_RECORDS,
+        )
+    }
+}

airbyte-integrations/connectors/destination-postgres/src/test-integration/kotlin/io/airbyte/integrations/destination/postgres/component/PostgresTestTableOperationsClient.kt

@@ -0,0 +1,257 @@
+/*
+ * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
+ */
+
+package io.airbyte.integrations.destination.postgres.component
+
+import io.airbyte.cdk.load.component.TestTableOperationsClient
+import io.airbyte.cdk.load.data.AirbyteValue
+import io.airbyte.cdk.load.schema.model.TableName
+import io.airbyte.cdk.load.util.Jsons
+import io.airbyte.integrations.destination.postgres.client.PostgresAirbyteClient
+import io.micronaut.context.annotation.Requires
+import jakarta.inject.Singleton
+import java.time.OffsetDateTime
+import java.time.ZoneOffset
+import java.time.format.DateTimeFormatter
+import javax.sql.DataSource
+
+@Requires(env = ["component"])
+@Singleton
+class PostgresTestTableOperationsClient(
+    private val dataSource: DataSource,
+    private val client: PostgresAirbyteClient,
+) : TestTableOperationsClient {
+    override suspend fun ping() {
+        dataSource.connection.use { connection ->
+            connection.createStatement().use { statement -> statement.executeQuery("SELECT 1") }
+        }
+    }
+
+    override suspend fun dropNamespace(namespace: String) {
+        dataSource.connection.use { connection ->
+            connection.createStatement().use { statement ->
+                statement.execute("DROP SCHEMA IF EXISTS \"$namespace\" CASCADE")
+            }
+        }
+    }
+
+    override suspend fun insertRecords(table: TableName, records: List<Map<String, AirbyteValue>>) {
+        if (records.isEmpty()) return
+
+        // Get column types from database to handle jsonb columns properly
+        val columnTypes = getColumnTypes(table)
+
+        // Get all unique columns from ALL records to handle sparse data (e.g., CDC deletion column)
+        val columns = records.flatMap { it.keys }.distinct().toList()
+        val columnNames = columns.joinToString(", ") { "\"$it\"" }
+        val placeholders = columns.indices.joinToString(", ") { "?" }
+
+        val sql =
+            """
+            INSERT INTO "${table.namespace}"."${table.name}" ($columnNames)
+            VALUES ($placeholders)
+        """
+
+        dataSource.connection.use { connection ->
+            connection.prepareStatement(sql).use { statement ->
+                for (record in records) {
+                    columns.forEachIndexed { index, column ->
+                        val value = record[column]
+                        val columnType = columnTypes[column]
+                        setParameterValue(statement, index + 1, value, columnType)
+                    }
+                    statement.addBatch()
+                }
+                statement.executeBatch()
+            }
+        }
+    }
+
+    private fun getColumnTypes(table: TableName): Map<String, String> {
+        val columnTypes = mutableMapOf<String, String>()
+        dataSource.connection.use { connection ->
+            connection.createStatement().use { statement ->
+                statement
+                    .executeQuery(
+                        """
+                    SELECT column_name, data_type
+                    FROM information_schema.columns
+                    WHERE table_schema = '${table.namespace}'
+                    AND table_name = '${table.name}'
+                """
+                    )
+                    .use { resultSet ->
+                        while (resultSet.next()) {
+                            columnTypes[resultSet.getString("column_name")] =
+                                resultSet.getString("data_type")
+                        }
+                    }
+            }
+        }
+        return columnTypes
+    }
+
+    private fun setParameterValue(
+        statement: java.sql.PreparedStatement,
+        index: Int,
+        value: AirbyteValue?,
+        columnType: String?
+    ) {
+        // If column is jsonb, serialize any value as JSON
+        if (columnType == "jsonb") {
+            if (value == null || value is io.airbyte.cdk.load.data.NullValue) {
+                statement.setNull(index, java.sql.Types.OTHER)
+            } else {
+                val pgObject = org.postgresql.util.PGobject()
+                pgObject.type = "jsonb"
+                pgObject.value = serializeToJson(value)
+                statement.setObject(index, pgObject)
+            }
+            return
+        }
+
+        when (value) {
+            null,
+            is io.airbyte.cdk.load.data.NullValue -> statement.setNull(index, java.sql.Types.NULL)
+            is io.airbyte.cdk.load.data.StringValue -> statement.setString(index, value.value)
+            is io.airbyte.cdk.load.data.IntegerValue ->
+                statement.setLong(index, value.value.toLong())
+            is io.airbyte.cdk.load.data.NumberValue -> statement.setBigDecimal(index, value.value)
+            is io.airbyte.cdk.load.data.BooleanValue -> statement.setBoolean(index, value.value)
+            is io.airbyte.cdk.load.data.TimestampWithTimezoneValue -> {
+                val offsetDateTime = OffsetDateTime.parse(value.value.toString())
+                statement.setObject(index, offsetDateTime)
+            }
+            is io.airbyte.cdk.load.data.TimestampWithoutTimezoneValue -> {
+                val localDateTime = java.time.LocalDateTime.parse(value.value.toString())
+                statement.setObject(index, localDateTime)
+            }
+            is io.airbyte.cdk.load.data.DateValue -> {
+                val localDate = java.time.LocalDate.parse(value.value.toString())
+                statement.setObject(index, localDate)
+            }
+            is io.airbyte.cdk.load.data.TimeWithTimezoneValue -> {
+                statement.setString(index, value.value.toString())
+            }
+            is io.airbyte.cdk.load.data.TimeWithoutTimezoneValue -> {
+                val localTime = java.time.LocalTime.parse(value.value.toString())
+                statement.setObject(index, localTime)
+            }
+            is io.airbyte.cdk.load.data.ObjectValue -> {
+                val pgObject = org.postgresql.util.PGobject()
+                pgObject.type = "jsonb"
+                pgObject.value = Jsons.writeValueAsString(value.values)
+                statement.setObject(index, pgObject)
+            }
+            is io.airbyte.cdk.load.data.ArrayValue -> {
+                val pgObject = org.postgresql.util.PGobject()
+                pgObject.type = "jsonb"
+                pgObject.value = Jsons.writeValueAsString(value.values)
+                statement.setObject(index, pgObject)
+            }
+            else -> {
+                // For unknown types, try to serialize as string
+                statement.setString(index, value.toString())
+            }
+        }
+    }
+
+    private fun serializeToJson(value: AirbyteValue): String {
+        return when (value) {
+            is io.airbyte.cdk.load.data.StringValue -> Jsons.writeValueAsString(value.value)
+            is io.airbyte.cdk.load.data.IntegerValue -> value.value.toString()
+            is io.airbyte.cdk.load.data.NumberValue -> value.value.toString()
+            is io.airbyte.cdk.load.data.BooleanValue -> value.value.toString()
+            is io.airbyte.cdk.load.data.ObjectValue -> Jsons.writeValueAsString(value.values)
+            is io.airbyte.cdk.load.data.ArrayValue -> Jsons.writeValueAsString(value.values)
+            is io.airbyte.cdk.load.data.NullValue -> "null"
+            else -> Jsons.writeValueAsString(value.toString())
+        }
+    }
+
+    override suspend fun readTable(table: TableName): List<Map<String, Any>> {
+        dataSource.connection.use { connection ->
+            connection.createStatement().use { statement ->
+                statement
+                    .executeQuery("""SELECT * FROM "${table.namespace}"."${table.name}"""")
+                    .use { resultSet ->
+                        val metaData = resultSet.metaData
+                        val columnCount = metaData.columnCount
+                        val result = mutableListOf<Map<String, Any>>()
+
+                        while (resultSet.next()) {
+                            val row = mutableMapOf<String, Any>()
+                            for (i in 1..columnCount) {
+                                val columnName = metaData.getColumnName(i)
+                                val columnType = metaData.getColumnTypeName(i)
+                                when (columnType.lowercase()) {
+                                    "timestamptz" -> {
+                                        val value =
+                                            resultSet.getObject(i, OffsetDateTime::class.java)
+                                        if (value != null) {
+                                            val formattedTimestamp =
+                                                DateTimeFormatter.ISO_OFFSET_DATE_TIME.format(
+                                                    value.withOffsetSameInstant(ZoneOffset.UTC)
+                                                )
+                                            row[columnName] = formattedTimestamp
+                                        }
+                                    }
+                                    "timestamp" -> {
+                                        val value = resultSet.getTimestamp(i)
+                                        if (value != null) {
+                                            val localDateTime = value.toLocalDateTime()
+                                            row[columnName] =
+                                                DateTimeFormatter.ISO_LOCAL_DATE_TIME.format(
+                                                    localDateTime
+                                                )
+                                        }
+                                    }
+                                    "jsonb",
+                                    "json" -> {
+                                        val stringValue: String? = resultSet.getString(i)
+                                        if (stringValue != null) {
+                                            val parsedValue =
+                                                Jsons.readValue(stringValue, Any::class.java)
+                                            val actualValue =
+                                                when (parsedValue) {
+                                                    is Int -> parsedValue.toLong()
+                                                    else -> parsedValue
+                                                }
+                                            row[columnName] = actualValue
+                                        }
+                                    }
+                                    else -> {
+                                        val value = resultSet.getObject(i)
+                                        if (value != null) {
+                                            // For varchar columns that may contain JSON (from
+                                            // schema evolution),
+                                            // normalize the JSON to compact format for comparison
+                                            if (
+                                                value is String &&
+                                                    (value.startsWith("{") || value.startsWith("["))
+                                            ) {
+                                                try {
+                                                    val parsed =
+                                                        Jsons.readValue(value, Any::class.java)
+                                                    row[columnName] =
+                                                        Jsons.writeValueAsString(parsed)
+                                                } catch (_: Exception) {
+                                                    row[columnName] = value
+                                                }
+                                            } else {
+                                                row[columnName] = value
+                                            }
+                                        }
+                                    }
+                                }
+                            }
+                            result.add(row)
+                        }
+
+                        return result
+                    }
+            }
+        }
+    }
+}

airbyte-integrations/connectors/destination-postgres/src/test-integration/kotlin/io/airbyte/integrations/destination/postgres/write/PostgresRawDataDumper.kt

@@ -267,7 +267,7 @@ class PostgresRawDataDumper(
                         .lowercase()
                         .toPostgresCompatibleName()
 
-                val fullyQualifiedTableName = "$rawNamespace.$rawName"
+                val fullyQualifiedTableName = "\"$rawNamespace\".\"$rawName\""
 
                 // Check if table exists first
                 val tableExistsQuery =
@@ -302,6 +302,26 @@ class PostgresRawDataDumper(
                         false
                     }
 
+                // Build the column name mapping from original names to transformed names
+                // We use the stream schema to get the original field names, then transform them
+                // using the postgres name transformation logic
+                val finalToInputColumnNames = mutableMapOf<String, String>()
+                if (stream.schema is ObjectType) {
+                    val objectSchema = stream.schema as ObjectType
+                    for (fieldName in objectSchema.properties.keys) {
+                        val transformedName = fieldName.toPostgresCompatibleName()
+                        // Map transformed name back to original name
+                        finalToInputColumnNames[transformedName] = fieldName
+                    }
+                }
+                // Also check if inputToFinalColumnNames mapping is available
+                val inputToFinalColumnNames =
+                    stream.tableSchema.columnSchema.inputToFinalColumnNames
+                // Add entries from the existing mapping (in case it was populated)
+                for ((input, final) in inputToFinalColumnNames) {
+                    finalToInputColumnNames[final] = input
+                }
+
                 while (resultSet.next()) {
                     val rawData =
                         if (hasDataColumn) {
@@ -313,8 +333,22 @@ class PostgresRawDataDumper(
                                     else -> dataObject?.toString() ?: "{}"
                                 }
 
-                            // Parse JSON to AirbyteValue, then coerce it to match the schema
-                            dataJson?.deserializeToNode()?.toAirbyteValue() ?: NullValue
+                            // Parse JSON to AirbyteValue, then map column names back to originals
+                            val parsedValue =
+                                dataJson?.deserializeToNode()?.toAirbyteValue() ?: NullValue
+                            // If the parsed value is an ObjectValue, map the column names back
+                            if (parsedValue is ObjectValue) {
+                                val mappedProperties = linkedMapOf<String, AirbyteValue>()
+                                for ((key, value) in parsedValue.values) {
+                                    // Map final column name back to input column name if mapping
+                                    // exists
+                                    val originalKey = finalToInputColumnNames[key] ?: key
+                                    mappedProperties[originalKey] = value
+                                }
+                                ObjectValue(mappedProperties)
+                            } else {
+                                parsedValue
+                            }
                         } else {
                             // Typed table mode: read from individual columns and reconstruct the
                             // object
@@ -333,10 +367,19 @@ class PostgresRawDataDumper(
 
                                 for ((fieldName, fieldType) in objectSchema.properties) {
                                     try {
+                                        // Map input field name to the transformed final column name
+                                        // First check the inputToFinalColumnNames mapping, then
+                                        // fall
+                                        // back to applying postgres transformation directly
+                                        val transformedColumnName =
+                                            inputToFinalColumnNames[fieldName]
+                                                ?: fieldName.toPostgresCompatibleName()
+
                                         // Try to find the actual column name (case-insensitive
                                         // lookup)
                                         val actualColumnName =
-                                            columnMap[fieldName.lowercase()] ?: fieldName
+                                            columnMap[transformedColumnName.lowercase()]
+                                                ?: transformedColumnName
                                         val columnValue = resultSet.getObject(actualColumnName)
                                         properties[fieldName] =
                                             when (columnValue) {

airbyte-integrations/connectors/destination-snowflake/gradle.properties

@@ -1,3 +1,3 @@
 testExecutionConcurrency=-1
-cdkVersion=0.1.82
+cdkVersion=0.1.91
 JunitMethodExecutionTimeout=10m

airbyte-integrations/connectors/destination-snowflake/metadata.yaml

@@ -6,7 +6,7 @@ data:
   connectorSubtype: database
   connectorType: destination
   definitionId: 424892c4-daac-4491-b35d-c6688ba547ba
-  dockerImageTag: 4.0.31
+  dockerImageTag: 4.0.32-rc.1
   dockerRepository: airbyte/destination-snowflake
   documentationUrl: https://docs.airbyte.com/integrations/destinations/snowflake
   githubIssueLabel: destination-snowflake
@@ -31,6 +31,8 @@ data:
       enabled: true
   releaseStage: generally_available
   releases:
+    rolloutConfiguration:
+      enableProgressiveRollout: true
     breakingChanges:
       2.0.0:
         message: Remove GCS/S3 loading method support.

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/SnowflakeBeanFactory.kt

@@ -11,15 +11,18 @@ import io.airbyte.cdk.load.check.CheckOperationV2
 import io.airbyte.cdk.load.check.DestinationCheckerV2
 import io.airbyte.cdk.load.config.DataChannelMedium
 import io.airbyte.cdk.load.dataflow.config.AggregatePublishingConfig
-import io.airbyte.cdk.load.orchestration.db.DefaultTempTableNameGenerator
-import io.airbyte.cdk.load.orchestration.db.TempTableNameGenerator
+import io.airbyte.cdk.load.table.DefaultTempTableNameGenerator
+import io.airbyte.cdk.load.table.TempTableNameGenerator
 import io.airbyte.cdk.output.OutputConsumer
 import io.airbyte.integrations.destination.snowflake.cdk.SnowflakeMigratingConfigurationSpecificationSupplier
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
+import io.airbyte.integrations.destination.snowflake.schema.toSnowflakeCompatibleName
 import io.airbyte.integrations.destination.snowflake.spec.KeyPairAuthConfiguration
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfigurationFactory
 import io.airbyte.integrations.destination.snowflake.spec.UsernamePasswordAuthConfiguration
+import io.airbyte.integrations.destination.snowflake.write.load.SnowflakeRawRecordFormatter
+import io.airbyte.integrations.destination.snowflake.write.load.SnowflakeRecordFormatter
+import io.airbyte.integrations.destination.snowflake.write.load.SnowflakeSchemaRecordFormatter
 import io.micronaut.context.annotation.Factory
 import io.micronaut.context.annotation.Primary
 import io.micronaut.context.annotation.Requires
@@ -204,6 +207,17 @@ class SnowflakeBeanFactory {
         outputConsumer: OutputConsumer,
     ) = CheckOperationV2(destinationChecker, outputConsumer)
 
+    @Singleton
+    fun snowflakeRecordFormatter(
+        snowflakeConfiguration: SnowflakeConfiguration
+    ): SnowflakeRecordFormatter {
+        return if (snowflakeConfiguration.legacyRawTablesOnly) {
+            SnowflakeRawRecordFormatter()
+        } else {
+            SnowflakeSchemaRecordFormatter()
+        }
+    }
+
     @Singleton
     fun aggregatePublishingConfig(dataChannelMedium: DataChannelMedium): AggregatePublishingConfig {
         // NOT speed mode

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/check/SnowflakeChecker.kt

@@ -13,13 +13,17 @@ import io.airbyte.cdk.load.data.FieldType
 import io.airbyte.cdk.load.data.ObjectType
 import io.airbyte.cdk.load.data.StringType
 import io.airbyte.cdk.load.message.Meta
+import io.airbyte.cdk.load.schema.model.ColumnSchema
+import io.airbyte.cdk.load.schema.model.StreamTableSchema
+import io.airbyte.cdk.load.schema.model.TableName
+import io.airbyte.cdk.load.schema.model.TableNames
 import io.airbyte.cdk.load.table.ColumnNameMapping
-import io.airbyte.cdk.load.table.TableName
 import io.airbyte.integrations.destination.snowflake.client.SnowflakeAirbyteClient
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
+import io.airbyte.integrations.destination.snowflake.schema.SnowflakeColumnManager
+import io.airbyte.integrations.destination.snowflake.schema.toSnowflakeCompatibleName
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
-import io.airbyte.integrations.destination.snowflake.sql.SnowflakeColumnUtils
 import io.airbyte.integrations.destination.snowflake.write.load.SnowflakeInsertBuffer
+import io.airbyte.integrations.destination.snowflake.write.load.SnowflakeSchemaRecordFormatter
 import jakarta.inject.Singleton
 import java.time.OffsetDateTime
 import java.util.UUID
@@ -31,7 +35,7 @@ internal const val CHECK_COLUMN_NAME = "test_key"
 class SnowflakeChecker(
     private val snowflakeAirbyteClient: SnowflakeAirbyteClient,
     private val snowflakeConfiguration: SnowflakeConfiguration,
-    private val snowflakeColumnUtils: SnowflakeColumnUtils,
+    private val columnManager: SnowflakeColumnManager,
 ) : DestinationCheckerV2 {
 
     override fun check() {
@@ -46,11 +50,40 @@ class SnowflakeChecker(
                 Meta.AirbyteMetaFields.GENERATION_ID.fieldName to AirbyteValue.from(0),
                 CHECK_COLUMN_NAME.toSnowflakeCompatibleName() to AirbyteValue.from("test-value")
             )
-        val outputSchema = snowflakeConfiguration.schema.toSnowflakeCompatibleName()
+        val outputSchema =
+            if (snowflakeConfiguration.legacyRawTablesOnly) {
+                snowflakeConfiguration.schema
+            } else {
+                snowflakeConfiguration.schema.toSnowflakeCompatibleName()
+            }
         val tableName =
             "_airbyte_connection_test_${
                 UUID.randomUUID().toString().replace("-".toRegex(), "")}".toSnowflakeCompatibleName()
         val qualifiedTableName = TableName(namespace = outputSchema, name = tableName)
+        val tableSchema =
+            StreamTableSchema(
+                tableNames =
+                    TableNames(
+                        finalTableName = qualifiedTableName,
+                        tempTableName = qualifiedTableName
+                    ),
+                columnSchema =
+                    ColumnSchema(
+                        inputToFinalColumnNames =
+                            mapOf(
+                                CHECK_COLUMN_NAME to CHECK_COLUMN_NAME.toSnowflakeCompatibleName()
+                            ),
+                        finalSchema =
+                            mapOf(
+                                CHECK_COLUMN_NAME.toSnowflakeCompatibleName() to
+                                    io.airbyte.cdk.load.component.ColumnType("VARCHAR", false)
+                            ),
+                        inputSchema =
+                            mapOf(CHECK_COLUMN_NAME to FieldType(StringType, nullable = false))
+                    ),
+                importType = Append
+            )
+
         val destinationStream =
             DestinationStream(
                 unmappedNamespace = outputSchema,
@@ -63,7 +96,8 @@ class SnowflakeChecker(
                 generationId = 0L,
                 minimumGenerationId = 0L,
                 syncId = 0L,
-                namespaceMapper = NamespaceMapper()
+                namespaceMapper = NamespaceMapper(),
+                tableSchema = tableSchema
             )
         runBlocking {
             try {
@@ -75,14 +109,14 @@ class SnowflakeChecker(
                     replace = true,
                 )
 
-                val columns = snowflakeAirbyteClient.describeTable(qualifiedTableName)
                 val snowflakeInsertBuffer =
                     SnowflakeInsertBuffer(
                         tableName = qualifiedTableName,
-                        columns = columns,
                         snowflakeClient = snowflakeAirbyteClient,
                         snowflakeConfiguration = snowflakeConfiguration,
-                        snowflakeColumnUtils = snowflakeColumnUtils,
+                        columnSchema = tableSchema.columnSchema,
+                        columnManager = columnManager,
+                        snowflakeRecordFormatter = SnowflakeSchemaRecordFormatter(),
                     )
 
                 snowflakeInsertBuffer.accumulate(data)

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/client/SnowflakeAirbyteClient.kt

@@ -13,18 +13,16 @@ import io.airbyte.cdk.load.component.TableColumns
 import io.airbyte.cdk.load.component.TableOperationsClient
 import io.airbyte.cdk.load.component.TableSchema
 import io.airbyte.cdk.load.component.TableSchemaEvolutionClient
+import io.airbyte.cdk.load.schema.model.TableName
 import io.airbyte.cdk.load.table.ColumnNameMapping
-import io.airbyte.cdk.load.table.TableName
 import io.airbyte.cdk.load.util.deserializeToNode
-import io.airbyte.integrations.destination.snowflake.db.ColumnDefinition
-import io.airbyte.integrations.destination.snowflake.db.escapeJsonIdentifier
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
+import io.airbyte.integrations.destination.snowflake.schema.SnowflakeColumnManager
+import io.airbyte.integrations.destination.snowflake.schema.toSnowflakeCompatibleName
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
 import io.airbyte.integrations.destination.snowflake.sql.COUNT_TOTAL_ALIAS
-import io.airbyte.integrations.destination.snowflake.sql.NOT_NULL
-import io.airbyte.integrations.destination.snowflake.sql.SnowflakeColumnUtils
 import io.airbyte.integrations.destination.snowflake.sql.SnowflakeDirectLoadSqlGenerator
 import io.airbyte.integrations.destination.snowflake.sql.andLog
+import io.airbyte.integrations.destination.snowflake.sql.escapeJsonIdentifier
 import io.github.oshai.kotlinlogging.KotlinLogging
 import jakarta.inject.Singleton
 import java.sql.ResultSet
@@ -41,13 +39,10 @@ private val log = KotlinLogging.logger {}
 class SnowflakeAirbyteClient(
     private val dataSource: DataSource,
     private val sqlGenerator: SnowflakeDirectLoadSqlGenerator,
-    private val snowflakeColumnUtils: SnowflakeColumnUtils,
     private val snowflakeConfiguration: SnowflakeConfiguration,
+    private val columnManager: SnowflakeColumnManager,
 ) : TableOperationsClient, TableSchemaEvolutionClient {
 
-    private val airbyteColumnNames =
-        snowflakeColumnUtils.getFormattedDefaultColumnNames(false).toSet()
-
     override suspend fun countTable(tableName: TableName): Long? =
         try {
             dataSource.connection.use { connection ->
@@ -126,7 +121,7 @@ class SnowflakeAirbyteClient(
         columnNameMapping: ColumnNameMapping,
         replace: Boolean
     ) {
-        execute(sqlGenerator.createTable(stream, tableName, columnNameMapping, replace))
+        execute(sqlGenerator.createTable(tableName, stream.tableSchema, replace))
         execute(sqlGenerator.createSnowflakeStage(tableName))
     }
 
@@ -163,7 +158,15 @@ class SnowflakeAirbyteClient(
         sourceTableName: TableName,
         targetTableName: TableName
     ) {
-        execute(sqlGenerator.copyTable(columnNameMapping, sourceTableName, targetTableName))
+        // Get all column names from the mapping (both meta columns and user columns)
+        val columnNames = buildSet {
+            // Add Airbyte meta columns (using uppercase constants)
+            addAll(columnManager.getMetaColumnNames())
+            // Add user columns from mapping
+            addAll(columnNameMapping.values)
+        }
+
+        execute(sqlGenerator.copyTable(columnNames, sourceTableName, targetTableName))
     }
 
     override suspend fun upsertTable(
@@ -172,9 +175,7 @@ class SnowflakeAirbyteClient(
         sourceTableName: TableName,
         targetTableName: TableName
     ) {
-        execute(
-            sqlGenerator.upsertTable(stream, columnNameMapping, sourceTableName, targetTableName)
-        )
+        execute(sqlGenerator.upsertTable(stream.tableSchema, sourceTableName, targetTableName))
     }
 
     override suspend fun dropTable(tableName: TableName) {
@@ -206,7 +207,7 @@ class SnowflakeAirbyteClient(
         stream: DestinationStream,
         columnNameMapping: ColumnNameMapping
     ): TableSchema {
-        return TableSchema(getColumnsFromStream(stream, columnNameMapping))
+        return TableSchema(stream.tableSchema.columnSchema.finalSchema)
     }
 
     override suspend fun applyChangeset(
@@ -253,7 +254,7 @@ class SnowflakeAirbyteClient(
                         val columnName = escapeJsonIdentifier(rs.getString("name"))
 
                         // Filter out airbyte columns
-                        if (airbyteColumnNames.contains(columnName)) {
+                        if (columnManager.getMetaColumnNames().contains(columnName)) {
                             continue
                         }
                         val dataType = rs.getString("type").takeWhile { char -> char != '(' }
@@ -271,49 +272,6 @@ class SnowflakeAirbyteClient(
         }
     }
 
-    internal fun getColumnsFromStream(
-        stream: DestinationStream,
-        columnNameMapping: ColumnNameMapping
-    ): Map<String, ColumnType> =
-        snowflakeColumnUtils
-            .columnsAndTypes(stream.schema.asColumns(), columnNameMapping)
-            .filter { column -> column.columnName !in airbyteColumnNames }
-            .associate { column ->
-                // columnsAndTypes returns types as either `FOO` or `FOO NOT NULL`.
-                // so check for that suffix.
-                val nullable = !column.columnType.endsWith(NOT_NULL)
-                val type =
-                    column.columnType
-                        .takeWhile { char ->
-                            // This is to remove any precision parts of the dialect type
-                            char != '('
-                        }
-                        .removeSuffix(NOT_NULL)
-                        .trim()
-
-                column.columnName to ColumnType(type, nullable)
-            }
-
-    internal fun generateSchemaChanges(
-        columnsInDb: Set<ColumnDefinition>,
-        columnsInStream: Set<ColumnDefinition>
-    ): Triple<Set<ColumnDefinition>, Set<ColumnDefinition>, Set<ColumnDefinition>> {
-        val addedColumns =
-            columnsInStream.filter { it.name !in columnsInDb.map { col -> col.name } }.toSet()
-        val deletedColumns =
-            columnsInDb.filter { it.name !in columnsInStream.map { col -> col.name } }.toSet()
-        val commonColumns =
-            columnsInStream.filter { it.name in columnsInDb.map { col -> col.name } }.toSet()
-        val modifiedColumns =
-            commonColumns
-                .filter {
-                    val dbType = columnsInDb.find { column -> it.name == column.name }?.type
-                    it.type != dbType
-                }
-                .toSet()
-        return Triple(addedColumns, deletedColumns, modifiedColumns)
-    }
-
     override suspend fun getGenerationId(tableName: TableName): Long =
         try {
             dataSource.connection.use { connection ->
@@ -326,7 +284,7 @@ class SnowflakeAirbyteClient(
                          * format.  In order to make sure these strings will match any column names
                          * that we have formatted in-memory, re-apply the escaping.
                          */
-                        resultSet.getLong(snowflakeColumnUtils.getGenerationIdColumnName())
+                        resultSet.getLong(columnManager.getGenerationIdColumnName())
                     } else {
                         log.warn {
                             "No generation ID found for table ${tableName.toPrettyString()}, returning 0"
@@ -351,8 +309,8 @@ class SnowflakeAirbyteClient(
         execute(sqlGenerator.putInStage(tableName, tempFilePath))
     }
 
-    fun copyFromStage(tableName: TableName, filename: String) {
-        execute(sqlGenerator.copyFromStage(tableName, filename))
+    fun copyFromStage(tableName: TableName, filename: String, columnNames: List<String>) {
+        execute(sqlGenerator.copyFromStage(tableName, filename, columnNames))
     }
 
     fun describeTable(tableName: TableName): LinkedHashMap<String, String> =

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/dataflow/SnowflakeAggregateFactory.kt

@@ -4,47 +4,41 @@
 
 package io.airbyte.integrations.destination.snowflake.dataflow
 
+import io.airbyte.cdk.load.command.DestinationCatalog
 import io.airbyte.cdk.load.dataflow.aggregate.Aggregate
 import io.airbyte.cdk.load.dataflow.aggregate.AggregateFactory
 import io.airbyte.cdk.load.dataflow.aggregate.StoreKey
-import io.airbyte.cdk.load.orchestration.db.direct_load_table.DirectLoadTableExecutionConfig
-import io.airbyte.cdk.load.table.TableName
+import io.airbyte.cdk.load.table.directload.DirectLoadTableExecutionConfig
 import io.airbyte.cdk.load.write.StreamStateStore
 import io.airbyte.integrations.destination.snowflake.client.SnowflakeAirbyteClient
+import io.airbyte.integrations.destination.snowflake.schema.SnowflakeColumnManager
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
-import io.airbyte.integrations.destination.snowflake.sql.SnowflakeColumnUtils
 import io.airbyte.integrations.destination.snowflake.write.load.SnowflakeInsertBuffer
-import io.micronaut.cache.annotation.CacheConfig
-import io.micronaut.cache.annotation.Cacheable
+import io.airbyte.integrations.destination.snowflake.write.load.SnowflakeRecordFormatter
 import jakarta.inject.Singleton
 
 @Singleton
-@CacheConfig("table-columns")
-// class has to be open to make the cache stuff work
-open class SnowflakeAggregateFactory(
+class SnowflakeAggregateFactory(
     private val snowflakeClient: SnowflakeAirbyteClient,
     private val streamStateStore: StreamStateStore<DirectLoadTableExecutionConfig>,
     private val snowflakeConfiguration: SnowflakeConfiguration,
-    private val snowflakeColumnUtils: SnowflakeColumnUtils,
+    private val catalog: DestinationCatalog,
+    private val columnManager: SnowflakeColumnManager,
+    private val snowflakeRecordFormatter: SnowflakeRecordFormatter,
 ) : AggregateFactory {
     override fun create(key: StoreKey): Aggregate {
+        val stream = catalog.getStream(key)
         val tableName = streamStateStore.get(key)!!.tableName
+
         val buffer =
             SnowflakeInsertBuffer(
                 tableName = tableName,
-                columns = getTableColumns(tableName),
                 snowflakeClient = snowflakeClient,
                 snowflakeConfiguration = snowflakeConfiguration,
-                snowflakeColumnUtils = snowflakeColumnUtils,
+                columnSchema = stream.tableSchema.columnSchema,
+                columnManager = columnManager,
+                snowflakeRecordFormatter = snowflakeRecordFormatter,
             )
         return SnowflakeAggregate(buffer = buffer)
     }
-
-    // We assume that a table isn't getting altered _during_ a sync.
-    // This allows us to only SHOW COLUMNS once per table per sync,
-    // rather than refetching it on every aggregate.
-    @Cacheable
-    // function has to be open to make caching work
-    internal open fun getTableColumns(tableName: TableName) =
-        snowflakeClient.describeTable(tableName)
 }

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/db/ColumnDefinition.kt

@@ -1,13 +0,0 @@
-/*
- * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
- */
-
-package io.airbyte.integrations.destination.snowflake.db
-
-/**
- * Jdbc destination column definition representation
- *
- * @param name
- * @param type
- */
-data class ColumnDefinition(val name: String, val type: String)

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/db/SnowflakeDirectLoadDatabaseInitialStatusGatherer.kt

@@ -4,17 +4,17 @@
 
 package io.airbyte.integrations.destination.snowflake.db
 
+import io.airbyte.cdk.load.command.DestinationCatalog
 import io.airbyte.cdk.load.component.TableOperationsClient
-import io.airbyte.cdk.load.orchestration.db.BaseDirectLoadInitialStatusGatherer
-import io.airbyte.cdk.load.orchestration.db.TempTableNameGenerator
+import io.airbyte.cdk.load.table.BaseDirectLoadInitialStatusGatherer
 import jakarta.inject.Singleton
 
 @Singleton
 class SnowflakeDirectLoadDatabaseInitialStatusGatherer(
     tableOperationsClient: TableOperationsClient,
-    tempTableNameGenerator: TempTableNameGenerator,
+    catalog: DestinationCatalog,
 ) :
     BaseDirectLoadInitialStatusGatherer(
         tableOperationsClient,
-        tempTableNameGenerator,
+        catalog,
     )

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/db/SnowflakeNameGenerators.kt

@@ -1,95 +0,0 @@
-/*
- * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
- */
-
-package io.airbyte.integrations.destination.snowflake.db
-
-import io.airbyte.cdk.ConfigErrorException
-import io.airbyte.cdk.load.command.DestinationStream
-import io.airbyte.cdk.load.orchestration.db.ColumnNameGenerator
-import io.airbyte.cdk.load.orchestration.db.FinalTableNameGenerator
-import io.airbyte.cdk.load.orchestration.db.legacy_typing_deduping.TypingDedupingUtil
-import io.airbyte.cdk.load.table.TableName
-import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
-import io.airbyte.integrations.destination.snowflake.sql.QUOTE
-import jakarta.inject.Singleton
-
-@Singleton
-class SnowflakeFinalTableNameGenerator(private val config: SnowflakeConfiguration) :
-    FinalTableNameGenerator {
-    override fun getTableName(streamDescriptor: DestinationStream.Descriptor): TableName {
-        val namespace = streamDescriptor.namespace ?: config.schema
-        return if (!config.legacyRawTablesOnly) {
-            TableName(
-                namespace = namespace.toSnowflakeCompatibleName(),
-                name = streamDescriptor.name.toSnowflakeCompatibleName(),
-            )
-        } else {
-            TableName(
-                namespace = config.internalTableSchema,
-                name =
-                    TypingDedupingUtil.concatenateRawTableName(
-                        namespace = escapeJsonIdentifier(namespace),
-                        name = escapeJsonIdentifier(streamDescriptor.name),
-                    ),
-            )
-        }
-    }
-}
-
-@Singleton
-class SnowflakeColumnNameGenerator(private val config: SnowflakeConfiguration) :
-    ColumnNameGenerator {
-    override fun getColumnName(column: String): ColumnNameGenerator.ColumnName {
-        return if (!config.legacyRawTablesOnly) {
-            ColumnNameGenerator.ColumnName(
-                column.toSnowflakeCompatibleName(),
-                column.toSnowflakeCompatibleName(),
-            )
-        } else {
-            ColumnNameGenerator.ColumnName(
-                column,
-                column,
-            )
-        }
-    }
-}
-
-/**
- * Escapes double-quotes in a JSON identifier by doubling them. This shit is legacy -- I don't know
- * why this would be necessary but no harm in keeping it so I am keeping it.
- *
- * @return The escaped identifier.
- */
-fun escapeJsonIdentifier(identifier: String): String {
-    // Note that we don't need to escape backslashes here!
-    // The only special character in an identifier is the double-quote, which needs to be
-    // doubled.
-    return identifier.replace(QUOTE, "$QUOTE$QUOTE")
-}
-
-/**
- * Transforms a string to be compatible with Snowflake table and column names.
- *
- * @return The transformed string suitable for Snowflake identifiers.
- */
-fun String.toSnowflakeCompatibleName(): String {
-    var identifier = this
-
-    // Handle empty strings
-    if (identifier.isEmpty()) {
-        throw ConfigErrorException("Empty string is invalid identifier")
-    }
-
-    // Snowflake scripting language does something weird when the `${` bigram shows up in the
-    // script so replace these with something else.
-    // For completeness, if we trigger this, also replace closing curly braces with underscores.
-    if (identifier.contains("\${")) {
-        identifier = identifier.replace("$", "_").replace("{", "_").replace("}", "_")
-    }
-
-    // Escape double quotes
-    identifier = escapeJsonIdentifier(identifier)
-
-    return identifier.uppercase()
-}

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/schema/SnowflakeColumnManager.kt

@@ -0,0 +1,116 @@
+/*
+ * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
+ */
+
+package io.airbyte.integrations.destination.snowflake.schema
+
+import io.airbyte.cdk.load.component.ColumnType
+import io.airbyte.cdk.load.message.Meta
+import io.airbyte.cdk.load.schema.model.ColumnSchema
+import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
+import io.airbyte.integrations.destination.snowflake.sql.SNOWFLAKE_AB_EXTRACTED_AT
+import io.airbyte.integrations.destination.snowflake.sql.SNOWFLAKE_AB_GENERATION_ID
+import io.airbyte.integrations.destination.snowflake.sql.SNOWFLAKE_AB_META
+import io.airbyte.integrations.destination.snowflake.sql.SNOWFLAKE_AB_RAW_ID
+import io.airbyte.integrations.destination.snowflake.sql.SnowflakeDataType
+import jakarta.inject.Singleton
+
+/**
+ * Manages column names and ordering for Snowflake tables based on whether legacy raw tables mode is
+ * enabled.
+ *
+ * TODO: We should add meta column munging and raw table support to the CDK, so this extra layer of
+ * management shouldn't be necessary.
+ */
+@Singleton
+class SnowflakeColumnManager(
+    private val config: SnowflakeConfiguration,
+) {
+    /**
+     * Get the list of column names for a table in the order they should appear in the CSV file and
+     * COPY INTO statement.
+     *
+     * Warning: MUST match the order defined in SnowflakeRecordFormatter
+     *
+     * @param columnSchema The schema containing column information (ignored in raw mode)
+     * @return List of column names in the correct order
+     */
+    fun getTableColumnNames(columnSchema: ColumnSchema): List<String> {
+        return buildList {
+            addAll(getMetaColumnNames())
+            addAll(columnSchema.finalSchema.keys)
+        }
+    }
+
+    /**
+     * Get the list of Airbyte meta column names. In schema mode, these are uppercase. In raw mode,
+     * they are lowercase and included loaded_at
+     *
+     * @return Set of meta column names
+     */
+    fun getMetaColumnNames(): Set<String> =
+        if (config.legacyRawTablesOnly) {
+            Constants.rawModeMetaColNames
+        } else {
+            Constants.schemaModeMetaColNames
+        }
+
+    /**
+     * Get the Airbyte meta columns as a map of column name to ColumnType. This provides both the
+     * column names and their types for table creation.
+     *
+     * @param columnSchema The user column schema (used to check for CDC columns)
+     * @return Map of meta column names to their types
+     */
+    fun getMetaColumns(): LinkedHashMap<String, ColumnType> {
+        return if (config.legacyRawTablesOnly) {
+            Constants.rawModeMetaColumns
+        } else {
+            Constants.schemaModeMetaColumns
+        }
+    }
+
+    fun getGenerationIdColumnName(): String {
+        return if (config.legacyRawTablesOnly) {
+            Meta.COLUMN_NAME_AB_GENERATION_ID
+        } else {
+            SNOWFLAKE_AB_GENERATION_ID
+        }
+    }
+
+    object Constants {
+        val rawModeMetaColumns =
+            linkedMapOf(
+                Meta.COLUMN_NAME_AB_RAW_ID to ColumnType(SnowflakeDataType.VARCHAR.typeName, false),
+                Meta.COLUMN_NAME_AB_EXTRACTED_AT to
+                    ColumnType(
+                        SnowflakeDataType.TIMESTAMP_TZ.typeName,
+                        false,
+                    ),
+                Meta.COLUMN_NAME_AB_META to ColumnType(SnowflakeDataType.VARIANT.typeName, false),
+                Meta.COLUMN_NAME_AB_GENERATION_ID to
+                    ColumnType(
+                        SnowflakeDataType.NUMBER.typeName,
+                        true,
+                    ),
+                Meta.COLUMN_NAME_AB_LOADED_AT to
+                    ColumnType(
+                        SnowflakeDataType.TIMESTAMP_TZ.typeName,
+                        true,
+                    ),
+            )
+
+        val schemaModeMetaColumns =
+            linkedMapOf(
+                SNOWFLAKE_AB_RAW_ID to ColumnType(SnowflakeDataType.VARCHAR.typeName, false),
+                SNOWFLAKE_AB_EXTRACTED_AT to
+                    ColumnType(SnowflakeDataType.TIMESTAMP_TZ.typeName, false),
+                SNOWFLAKE_AB_META to ColumnType(SnowflakeDataType.VARIANT.typeName, false),
+                SNOWFLAKE_AB_GENERATION_ID to ColumnType(SnowflakeDataType.NUMBER.typeName, true),
+            )
+
+        val rawModeMetaColNames: Set<String> = rawModeMetaColumns.keys
+
+        val schemaModeMetaColNames: Set<String> = schemaModeMetaColumns.keys
+    }
+}

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/schema/SnowflakeNamingUtils.kt

@@ -0,0 +1,34 @@
+/*
+ * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
+ */
+
+package io.airbyte.integrations.destination.snowflake.schema
+
+import io.airbyte.cdk.ConfigErrorException
+import io.airbyte.integrations.destination.snowflake.sql.escapeJsonIdentifier
+
+/**
+ * Transforms a string to be compatible with Snowflake table and column names.
+ *
+ * @return The transformed string suitable for Snowflake identifiers.
+ */
+fun String.toSnowflakeCompatibleName(): String {
+    var identifier = this
+
+    // Handle empty strings
+    if (identifier.isEmpty()) {
+        throw ConfigErrorException("Empty string is invalid identifier")
+    }
+
+    // Snowflake scripting language does something weird when the `${` bigram shows up in the
+    // script so replace these with something else.
+    // For completeness, if we trigger this, also replace closing curly braces with underscores.
+    if (identifier.contains("\${")) {
+        identifier = identifier.replace("$", "_").replace("{", "_").replace("}", "_")
+    }
+
+    // Escape double quotes
+    identifier = escapeJsonIdentifier(identifier)
+
+    return identifier.uppercase()
+}

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/schema/SnowflakeTableSchemaMapper.kt

@@ -0,0 +1,121 @@
+/*
+ * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
+ */
+
+package io.airbyte.integrations.destination.snowflake.schema
+
+import io.airbyte.cdk.load.command.DestinationStream
+import io.airbyte.cdk.load.component.ColumnType
+import io.airbyte.cdk.load.data.ArrayType
+import io.airbyte.cdk.load.data.ArrayTypeWithoutSchema
+import io.airbyte.cdk.load.data.BooleanType
+import io.airbyte.cdk.load.data.DateType
+import io.airbyte.cdk.load.data.FieldType
+import io.airbyte.cdk.load.data.IntegerType
+import io.airbyte.cdk.load.data.NumberType
+import io.airbyte.cdk.load.data.ObjectType
+import io.airbyte.cdk.load.data.ObjectTypeWithEmptySchema
+import io.airbyte.cdk.load.data.ObjectTypeWithoutSchema
+import io.airbyte.cdk.load.data.StringType
+import io.airbyte.cdk.load.data.TimeTypeWithTimezone
+import io.airbyte.cdk.load.data.TimeTypeWithoutTimezone
+import io.airbyte.cdk.load.data.TimestampTypeWithTimezone
+import io.airbyte.cdk.load.data.TimestampTypeWithoutTimezone
+import io.airbyte.cdk.load.data.UnionType
+import io.airbyte.cdk.load.data.UnknownType
+import io.airbyte.cdk.load.message.Meta
+import io.airbyte.cdk.load.schema.TableSchemaMapper
+import io.airbyte.cdk.load.schema.model.StreamTableSchema
+import io.airbyte.cdk.load.schema.model.TableName
+import io.airbyte.cdk.load.table.TempTableNameGenerator
+import io.airbyte.cdk.load.table.TypingDedupingUtil
+import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
+import io.airbyte.integrations.destination.snowflake.sql.SnowflakeDataType
+import io.airbyte.integrations.destination.snowflake.sql.escapeJsonIdentifier
+import jakarta.inject.Singleton
+
+@Singleton
+class SnowflakeTableSchemaMapper(
+    private val config: SnowflakeConfiguration,
+    private val tempTableNameGenerator: TempTableNameGenerator,
+) : TableSchemaMapper {
+    override fun toFinalTableName(desc: DestinationStream.Descriptor): TableName {
+        val namespace = desc.namespace ?: config.schema
+        return if (!config.legacyRawTablesOnly) {
+            TableName(
+                namespace = namespace.toSnowflakeCompatibleName(),
+                name = desc.name.toSnowflakeCompatibleName(),
+            )
+        } else {
+            TableName(
+                namespace = config.internalTableSchema,
+                name =
+                    TypingDedupingUtil.concatenateRawTableName(
+                        namespace = escapeJsonIdentifier(namespace),
+                        name = escapeJsonIdentifier(desc.name),
+                    ),
+            )
+        }
+    }
+
+    override fun toTempTableName(tableName: TableName): TableName {
+        return tempTableNameGenerator.generate(tableName)
+    }
+
+    override fun toColumnName(name: String): String {
+        return if (!config.legacyRawTablesOnly) {
+            name.toSnowflakeCompatibleName()
+        } else {
+            // In legacy mode, column names are not transformed
+            name
+        }
+    }
+
+    override fun toColumnType(fieldType: FieldType): ColumnType {
+        val snowflakeType =
+            when (fieldType.type) {
+                // Simple types
+                BooleanType -> SnowflakeDataType.BOOLEAN.typeName
+                IntegerType -> SnowflakeDataType.NUMBER.typeName
+                NumberType -> SnowflakeDataType.FLOAT.typeName
+                StringType -> SnowflakeDataType.VARCHAR.typeName
+
+                // Temporal types
+                DateType -> SnowflakeDataType.DATE.typeName
+                TimeTypeWithTimezone -> SnowflakeDataType.VARCHAR.typeName
+                TimeTypeWithoutTimezone -> SnowflakeDataType.TIME.typeName
+                TimestampTypeWithTimezone -> SnowflakeDataType.TIMESTAMP_TZ.typeName
+                TimestampTypeWithoutTimezone -> SnowflakeDataType.TIMESTAMP_NTZ.typeName
+
+                // Semistructured types
+                is ArrayType,
+                ArrayTypeWithoutSchema -> SnowflakeDataType.ARRAY.typeName
+                is ObjectType,
+                ObjectTypeWithEmptySchema,
+                ObjectTypeWithoutSchema -> SnowflakeDataType.OBJECT.typeName
+                is UnionType -> SnowflakeDataType.VARIANT.typeName
+                is UnknownType -> SnowflakeDataType.VARIANT.typeName
+            }
+
+        return ColumnType(snowflakeType, fieldType.nullable)
+    }
+
+    override fun toFinalSchema(tableSchema: StreamTableSchema): StreamTableSchema {
+        if (!config.legacyRawTablesOnly) {
+            return tableSchema
+        }
+
+        return StreamTableSchema(
+            tableNames = tableSchema.tableNames,
+            columnSchema =
+                tableSchema.columnSchema.copy(
+                    finalSchema =
+                        mapOf(
+                            Meta.COLUMN_NAME_DATA to
+                                ColumnType(SnowflakeDataType.OBJECT.typeName, false)
+                        )
+                ),
+            importType = tableSchema.importType,
+        )
+    }
+}

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/sql/SnowflakeColumnUtils.kt

@@ -1,201 +0,0 @@
-/*
- * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
- */
-
-package io.airbyte.integrations.destination.snowflake.sql
-
-import com.google.common.annotations.VisibleForTesting
-import io.airbyte.cdk.load.data.AirbyteType
-import io.airbyte.cdk.load.data.ArrayType
-import io.airbyte.cdk.load.data.ArrayTypeWithoutSchema
-import io.airbyte.cdk.load.data.BooleanType
-import io.airbyte.cdk.load.data.DateType
-import io.airbyte.cdk.load.data.FieldType
-import io.airbyte.cdk.load.data.IntegerType
-import io.airbyte.cdk.load.data.NumberType
-import io.airbyte.cdk.load.data.ObjectType
-import io.airbyte.cdk.load.data.ObjectTypeWithEmptySchema
-import io.airbyte.cdk.load.data.ObjectTypeWithoutSchema
-import io.airbyte.cdk.load.data.StringType
-import io.airbyte.cdk.load.data.TimeTypeWithTimezone
-import io.airbyte.cdk.load.data.TimeTypeWithoutTimezone
-import io.airbyte.cdk.load.data.TimestampTypeWithTimezone
-import io.airbyte.cdk.load.data.TimestampTypeWithoutTimezone
-import io.airbyte.cdk.load.data.UnionType
-import io.airbyte.cdk.load.data.UnknownType
-import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_EXTRACTED_AT
-import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_GENERATION_ID
-import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_LOADED_AT
-import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_META
-import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_RAW_ID
-import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_DATA
-import io.airbyte.cdk.load.table.ColumnNameMapping
-import io.airbyte.integrations.destination.snowflake.db.SnowflakeColumnNameGenerator
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
-import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
-import jakarta.inject.Singleton
-import kotlin.collections.component1
-import kotlin.collections.component2
-import kotlin.collections.joinToString
-import kotlin.collections.map
-import kotlin.collections.plus
-
-internal const val NOT_NULL = "NOT NULL"
-
-internal val DEFAULT_COLUMNS =
-    listOf(
-        ColumnAndType(
-            columnName = COLUMN_NAME_AB_RAW_ID,
-            columnType = "${SnowflakeDataType.VARCHAR.typeName} $NOT_NULL"
-        ),
-        ColumnAndType(
-            columnName = COLUMN_NAME_AB_EXTRACTED_AT,
-            columnType = "${SnowflakeDataType.TIMESTAMP_TZ.typeName} $NOT_NULL"
-        ),
-        ColumnAndType(
-            columnName = COLUMN_NAME_AB_META,
-            columnType = "${SnowflakeDataType.VARIANT.typeName} $NOT_NULL"
-        ),
-        ColumnAndType(
-            columnName = COLUMN_NAME_AB_GENERATION_ID,
-            columnType = SnowflakeDataType.NUMBER.typeName
-        ),
-    )
-
-internal val RAW_DATA_COLUMN =
-    ColumnAndType(
-        columnName = COLUMN_NAME_DATA,
-        columnType = "${SnowflakeDataType.VARIANT.typeName} $NOT_NULL"
-    )
-
-internal val RAW_COLUMNS =
-    listOf(
-        ColumnAndType(
-            columnName = COLUMN_NAME_AB_LOADED_AT,
-            columnType = SnowflakeDataType.TIMESTAMP_TZ.typeName
-        ),
-        RAW_DATA_COLUMN
-    )
-
-@Singleton
-class SnowflakeColumnUtils(
-    private val snowflakeConfiguration: SnowflakeConfiguration,
-    private val snowflakeColumnNameGenerator: SnowflakeColumnNameGenerator,
-) {
-
-    @VisibleForTesting
-    internal fun defaultColumns(): List<ColumnAndType> =
-        if (snowflakeConfiguration.legacyRawTablesOnly) {
-            DEFAULT_COLUMNS + RAW_COLUMNS
-        } else {
-            DEFAULT_COLUMNS
-        }
-
-    internal fun formattedDefaultColumns(): List<ColumnAndType> =
-        defaultColumns().map {
-            ColumnAndType(
-                columnName = formatColumnName(it.columnName, false),
-                columnType = it.columnType,
-            )
-        }
-
-    fun getGenerationIdColumnName(): String {
-        return if (snowflakeConfiguration.legacyRawTablesOnly) {
-            COLUMN_NAME_AB_GENERATION_ID
-        } else {
-            COLUMN_NAME_AB_GENERATION_ID.toSnowflakeCompatibleName()
-        }
-    }
-
-    fun getColumnNames(columnNameMapping: ColumnNameMapping): String =
-        if (snowflakeConfiguration.legacyRawTablesOnly) {
-            getFormattedDefaultColumnNames(true).joinToString(",")
-        } else {
-            (getFormattedDefaultColumnNames(true) +
-                    columnNameMapping.map { (_, actualName) -> actualName.quote() })
-                .joinToString(",")
-        }
-
-    fun getFormattedDefaultColumnNames(quote: Boolean = false): List<String> =
-        defaultColumns().map { formatColumnName(it.columnName, quote) }
-
-    fun getFormattedColumnNames(
-        columns: Map<String, FieldType>,
-        columnNameMapping: ColumnNameMapping,
-        quote: Boolean = true,
-    ): List<String> =
-        if (snowflakeConfiguration.legacyRawTablesOnly) {
-            getFormattedDefaultColumnNames(quote)
-        } else {
-            getFormattedDefaultColumnNames(quote) +
-                columns.map { (fieldName, _) ->
-                    val columnName = columnNameMapping[fieldName] ?: fieldName
-                    if (quote) columnName.quote() else columnName
-                }
-        }
-
-    fun columnsAndTypes(
-        columns: Map<String, FieldType>,
-        columnNameMapping: ColumnNameMapping
-    ): List<ColumnAndType> =
-        if (snowflakeConfiguration.legacyRawTablesOnly) {
-            formattedDefaultColumns()
-        } else {
-            formattedDefaultColumns() +
-                columns.map { (fieldName, type) ->
-                    val columnName = columnNameMapping[fieldName] ?: fieldName
-                    val typeName = toDialectType(type.type)
-                    ColumnAndType(
-                        columnName = columnName,
-                        columnType = if (type.nullable) typeName else "$typeName $NOT_NULL",
-                    )
-                }
-        }
-
-    fun formatColumnName(
-        columnName: String,
-        quote: Boolean = true,
-    ): String {
-        val formattedColumnName =
-            if (columnName == COLUMN_NAME_DATA) columnName
-            else snowflakeColumnNameGenerator.getColumnName(columnName).displayName
-        return if (quote) formattedColumnName.quote() else formattedColumnName
-    }
-
-    fun toDialectType(type: AirbyteType): String =
-        when (type) {
-            // Simple types
-            BooleanType -> SnowflakeDataType.BOOLEAN.typeName
-            IntegerType -> SnowflakeDataType.NUMBER.typeName
-            NumberType -> SnowflakeDataType.FLOAT.typeName
-            StringType -> SnowflakeDataType.VARCHAR.typeName
-
-            // Temporal types
-            DateType -> SnowflakeDataType.DATE.typeName
-            TimeTypeWithTimezone -> SnowflakeDataType.VARCHAR.typeName
-            TimeTypeWithoutTimezone -> SnowflakeDataType.TIME.typeName
-            TimestampTypeWithTimezone -> SnowflakeDataType.TIMESTAMP_TZ.typeName
-            TimestampTypeWithoutTimezone -> SnowflakeDataType.TIMESTAMP_NTZ.typeName
-
-            // Semistructured types
-            is ArrayType,
-            ArrayTypeWithoutSchema -> SnowflakeDataType.ARRAY.typeName
-            is ObjectType,
-            ObjectTypeWithEmptySchema,
-            ObjectTypeWithoutSchema -> SnowflakeDataType.OBJECT.typeName
-            is UnionType -> SnowflakeDataType.VARIANT.typeName
-            is UnknownType -> SnowflakeDataType.VARIANT.typeName
-        }
-}
-
-data class ColumnAndType(val columnName: String, val columnType: String) {
-    override fun toString(): String {
-        return "${columnName.quote()} $columnType"
-    }
-}
-
-/**
- * Surrounds the string instance with double quotation marks (e.g. "some string" -> "\"some
- * string\"").
- */
-fun String.quote() = "$QUOTE$this$QUOTE"

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/sql/SnowflakeDataType.kt

@@ -10,7 +10,7 @@ package io.airbyte.integrations.destination.snowflake.sql
  */
 enum class SnowflakeDataType(val typeName: String) {
     // Numeric types
-    NUMBER("NUMBER(38,0)"),
+    NUMBER("NUMBER"),
     FLOAT("FLOAT"),
 
     // String & binary types

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/sql/SnowflakeDirectLoadSqlGenerator.kt

@@ -4,16 +4,19 @@
 
 package io.airbyte.integrations.destination.snowflake.sql
 
-import io.airbyte.cdk.load.command.Dedupe
-import io.airbyte.cdk.load.command.DestinationStream
+import com.google.common.annotations.VisibleForTesting
 import io.airbyte.cdk.load.component.ColumnType
 import io.airbyte.cdk.load.component.ColumnTypeChange
 import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_EXTRACTED_AT
+import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_GENERATION_ID
+import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_META
+import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_RAW_ID
+import io.airbyte.cdk.load.schema.model.StreamTableSchema
+import io.airbyte.cdk.load.schema.model.TableName
 import io.airbyte.cdk.load.table.CDC_DELETED_AT_COLUMN
-import io.airbyte.cdk.load.table.ColumnNameMapping
-import io.airbyte.cdk.load.table.TableName
 import io.airbyte.cdk.load.util.UUIDGenerator
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
+import io.airbyte.integrations.destination.snowflake.schema.SnowflakeColumnManager
+import io.airbyte.integrations.destination.snowflake.schema.toSnowflakeCompatibleName
 import io.airbyte.integrations.destination.snowflake.spec.CdcDeletionMode
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
 import io.airbyte.integrations.destination.snowflake.write.load.CSV_FIELD_SEPARATOR
@@ -22,6 +25,14 @@ import io.github.oshai.kotlinlogging.KotlinLogging
 import jakarta.inject.Singleton
 
 internal const val COUNT_TOTAL_ALIAS = "TOTAL"
+internal const val NOT_NULL = "NOT NULL"
+
+// Snowflake-compatible (uppercase) versions of the Airbyte meta column names
+internal val SNOWFLAKE_AB_RAW_ID = COLUMN_NAME_AB_RAW_ID.toSnowflakeCompatibleName()
+internal val SNOWFLAKE_AB_EXTRACTED_AT = COLUMN_NAME_AB_EXTRACTED_AT.toSnowflakeCompatibleName()
+internal val SNOWFLAKE_AB_META = COLUMN_NAME_AB_META.toSnowflakeCompatibleName()
+internal val SNOWFLAKE_AB_GENERATION_ID = COLUMN_NAME_AB_GENERATION_ID.toSnowflakeCompatibleName()
+internal val SNOWFLAKE_AB_CDC_DELETED_AT_COLUMN = CDC_DELETED_AT_COLUMN.toSnowflakeCompatibleName()
 
 private val log = KotlinLogging.logger {}
 
@@ -36,80 +47,91 @@ fun String.andLog(): String {
 
 @Singleton
 class SnowflakeDirectLoadSqlGenerator(
-    private val columnUtils: SnowflakeColumnUtils,
     private val uuidGenerator: UUIDGenerator,
-    private val snowflakeConfiguration: SnowflakeConfiguration,
-    private val snowflakeSqlNameUtils: SnowflakeSqlNameUtils,
+    private val config: SnowflakeConfiguration,
+    private val columnManager: SnowflakeColumnManager,
 ) {
     fun countTable(tableName: TableName): String {
-        return "SELECT COUNT(*) AS $COUNT_TOTAL_ALIAS FROM ${snowflakeSqlNameUtils.fullyQualifiedName(tableName)}".andLog()
+        return "SELECT COUNT(*) AS $COUNT_TOTAL_ALIAS FROM ${fullyQualifiedName(tableName)}".andLog()
     }
 
     fun createNamespace(namespace: String): String {
-        return "CREATE SCHEMA IF NOT EXISTS ${snowflakeSqlNameUtils.fullyQualifiedNamespace(namespace)}".andLog()
+        return "CREATE SCHEMA IF NOT EXISTS ${fullyQualifiedNamespace(namespace)}".andLog()
     }
 
     fun createTable(
-        stream: DestinationStream,
         tableName: TableName,
-        columnNameMapping: ColumnNameMapping,
+        tableSchema: StreamTableSchema,
         replace: Boolean
     ): String {
+        val finalSchema = tableSchema.columnSchema.finalSchema
+        val metaColumns = columnManager.getMetaColumns()
+
+        // Build column declarations from the meta columns and user schema
         val columnDeclarations =
-            columnUtils
-                .columnsAndTypes(stream.schema.asColumns(), columnNameMapping)
-                .joinToString(",\n")
+            buildList {
+                    // Add Airbyte meta columns from the column manager
+                    metaColumns.forEach { (columnName, columnType) ->
+                        val nullability = if (columnType.nullable) "" else " NOT NULL"
+                        add("${columnName.quote()} ${columnType.type}$nullability")
+                    }
+
+                    // Add user columns from the munged schema
+                    finalSchema.forEach { (columnName, columnType) ->
+                        val nullability = if (columnType.nullable) "" else " NOT NULL"
+                        add("${columnName.quote()} ${columnType.type}$nullability")
+                    }
+                }
+                .joinToString(",\n    ")
 
         // Snowflake supports CREATE OR REPLACE TABLE, which is simpler than drop+recreate
         val createOrReplace = if (replace) "CREATE OR REPLACE" else "CREATE"
 
         val createTableStatement =
             """
-            $createOrReplace TABLE ${snowflakeSqlNameUtils.fullyQualifiedName(tableName)} (
-                $columnDeclarations
-            )
-        """.trimIndent()
+            |$createOrReplace TABLE ${fullyQualifiedName(tableName)} (
+            |    $columnDeclarations
+            |)
+            """.trimMargin() // Something was tripping up trimIndent so we opt for trimMargin
 
         return createTableStatement.andLog()
     }
 
     fun showColumns(tableName: TableName): String =
-        "SHOW COLUMNS IN TABLE ${snowflakeSqlNameUtils.fullyQualifiedName(tableName)}".andLog()
+        "SHOW COLUMNS IN TABLE ${fullyQualifiedName(tableName)}".andLog()
 
     fun copyTable(
-        columnNameMapping: ColumnNameMapping,
+        columnNames: Set<String>,
         sourceTableName: TableName,
         targetTableName: TableName
     ): String {
-        val columnNames = columnUtils.getColumnNames(columnNameMapping)
+        val columnList = columnNames.joinToString(", ") { it.quote() }
 
         return """
-            INSERT INTO ${snowflakeSqlNameUtils.fullyQualifiedName(targetTableName)} 
+            INSERT INTO ${fullyQualifiedName(targetTableName)} 
             (
-                $columnNames
+                $columnList
             )
             SELECT
-                $columnNames
-            FROM ${snowflakeSqlNameUtils.fullyQualifiedName(sourceTableName)}
+                $columnList
+            FROM ${fullyQualifiedName(sourceTableName)}
             """
             .trimIndent()
             .andLog()
     }
 
     fun upsertTable(
-        stream: DestinationStream,
-        columnNameMapping: ColumnNameMapping,
+        tableSchema: StreamTableSchema,
         sourceTableName: TableName,
         targetTableName: TableName
     ): String {
-        val importType = stream.importType as Dedupe
+        val finalSchema = tableSchema.columnSchema.finalSchema
 
         // Build primary key matching condition
+        val pks = tableSchema.getPrimaryKey().flatten()
         val pkEquivalent =
-            if (importType.primaryKey.isNotEmpty()) {
-                importType.primaryKey.joinToString(" AND ") { fieldPath ->
-                    val fieldName = fieldPath.first()
-                    val columnName = columnNameMapping[fieldName] ?: fieldName
+            if (pks.isNotEmpty()) {
+                pks.joinToString(" AND ") { columnName ->
                     val targetTableColumnName = "target_table.${columnName.quote()}"
                     val newRecordColumnName = "new_record.${columnName.quote()}"
                     """($targetTableColumnName = $newRecordColumnName OR ($targetTableColumnName IS NULL AND $newRecordColumnName IS NULL))"""
@@ -120,80 +142,62 @@ class SnowflakeDirectLoadSqlGenerator(
             }
 
         // Build column lists for INSERT and UPDATE
-        val columnList: String =
-            columnUtils
-                .getFormattedColumnNames(
-                    columns = stream.schema.asColumns(),
-                    columnNameMapping = columnNameMapping,
-                    quote = false,
-                )
-                .joinToString(
-                    ",\n",
-                ) {
-                    it.quote()
-                }
+        val allColumns = buildList {
+            add(SNOWFLAKE_AB_RAW_ID)
+            add(SNOWFLAKE_AB_EXTRACTED_AT)
+            add(SNOWFLAKE_AB_META)
+            add(SNOWFLAKE_AB_GENERATION_ID)
+            addAll(finalSchema.keys)
+        }
 
+        val columnList: String = allColumns.joinToString(",\n  ") { it.quote() }
         val newRecordColumnList: String =
-            columnUtils
-                .getFormattedColumnNames(
-                    columns = stream.schema.asColumns(),
-                    columnNameMapping = columnNameMapping,
-                    quote = false,
-                )
-                .joinToString(",\n") { "new_record.${it.quote()}" }
+            allColumns.joinToString(",\n  ") { "new_record.${it.quote()}" }
 
         // Get deduped records from source
-        val selectSourceRecords = selectDedupedRecords(stream, sourceTableName, columnNameMapping)
+        val selectSourceRecords = selectDedupedRecords(tableSchema, sourceTableName)
 
         // Build cursor comparison for determining which record is newer
         val cursorComparison: String
-        if (importType.cursor.isNotEmpty()) {
-            val cursorFieldName = importType.cursor.first()
-            val cursor = (columnNameMapping[cursorFieldName] ?: cursorFieldName)
+        val cursor = tableSchema.getCursor().firstOrNull()
+        if (cursor != null) {
             val targetTableCursor = "target_table.${cursor.quote()}"
             val newRecordCursor = "new_record.${cursor.quote()}"
             cursorComparison =
                 """
                 (
                   $targetTableCursor < $newRecordCursor
-                  OR ($targetTableCursor = $newRecordCursor AND target_table."${COLUMN_NAME_AB_EXTRACTED_AT.toSnowflakeCompatibleName()}" < new_record."${COLUMN_NAME_AB_EXTRACTED_AT.toSnowflakeCompatibleName()}")
-                  OR ($targetTableCursor IS NULL AND $newRecordCursor IS NULL AND target_table."${COLUMN_NAME_AB_EXTRACTED_AT.toSnowflakeCompatibleName()}" < new_record."${COLUMN_NAME_AB_EXTRACTED_AT.toSnowflakeCompatibleName()}")
+                  OR ($targetTableCursor = $newRecordCursor AND target_table."$SNOWFLAKE_AB_EXTRACTED_AT" < new_record."$SNOWFLAKE_AB_EXTRACTED_AT")
+                  OR ($targetTableCursor IS NULL AND $newRecordCursor IS NULL AND target_table."$SNOWFLAKE_AB_EXTRACTED_AT" < new_record."$SNOWFLAKE_AB_EXTRACTED_AT")
                   OR ($targetTableCursor IS NULL AND $newRecordCursor IS $NOT_NULL)
                 )
             """.trimIndent()
         } else {
             // No cursor - use extraction timestamp only
             cursorComparison =
-                """target_table."${COLUMN_NAME_AB_EXTRACTED_AT.toSnowflakeCompatibleName()}" < new_record."${COLUMN_NAME_AB_EXTRACTED_AT.toSnowflakeCompatibleName()}""""
+                """target_table."$SNOWFLAKE_AB_EXTRACTED_AT" < new_record."$SNOWFLAKE_AB_EXTRACTED_AT""""
         }
 
         // Build column assignments for UPDATE
         val columnAssignments: String =
-            columnUtils
-                .getFormattedColumnNames(
-                    columns = stream.schema.asColumns(),
-                    columnNameMapping = columnNameMapping,
-                    quote = false,
-                )
-                .joinToString(",\n") { column ->
-                    "${column.quote()} = new_record.${column.quote()}"
-                }
+            allColumns.joinToString(",\n  ") { column ->
+                "${column.quote()} = new_record.${column.quote()}"
+            }
 
         // Handle CDC deletions based on mode
         val cdcDeleteClause: String
         val cdcSkipInsertClause: String
         if (
-            stream.schema.asColumns().containsKey(CDC_DELETED_AT_COLUMN) &&
-                snowflakeConfiguration.cdcDeletionMode == CdcDeletionMode.HARD_DELETE
+            finalSchema.containsKey(SNOWFLAKE_AB_CDC_DELETED_AT_COLUMN) &&
+                config.cdcDeletionMode == CdcDeletionMode.HARD_DELETE
         ) {
             // Execute CDC deletions if there's already a record
             cdcDeleteClause =
-                "WHEN MATCHED AND new_record.\"${CDC_DELETED_AT_COLUMN.toSnowflakeCompatibleName()}\" IS NOT NULL AND $cursorComparison THEN DELETE"
+                "WHEN MATCHED AND new_record.\"${SNOWFLAKE_AB_CDC_DELETED_AT_COLUMN}\" IS NOT NULL AND $cursorComparison THEN DELETE"
             // And skip insertion entirely if there's no matching record.
             // (This is possible if a single T+D batch contains both an insertion and deletion for
             // the same PK)
-            cdcSkipInsertClause =
-                "AND new_record.\"${CDC_DELETED_AT_COLUMN.toSnowflakeCompatibleName()}\" IS NULL"
+            cdcSkipInsertClause = "AND new_record.\"${SNOWFLAKE_AB_CDC_DELETED_AT_COLUMN}\" IS NULL"
         } else {
             cdcDeleteClause = ""
             cdcSkipInsertClause = ""
@@ -203,35 +207,35 @@ class SnowflakeDirectLoadSqlGenerator(
         val mergeStatement =
             if (cdcDeleteClause.isNotEmpty()) {
                 """
-            MERGE INTO ${snowflakeSqlNameUtils.fullyQualifiedName(targetTableName)} AS target_table
-            USING (
-              $selectSourceRecords
-            ) AS new_record
-            ON $pkEquivalent
-            $cdcDeleteClause
-            WHEN MATCHED AND $cursorComparison THEN UPDATE SET
-              $columnAssignments
-            WHEN NOT MATCHED $cdcSkipInsertClause THEN INSERT (
-              $columnList
-            ) VALUES (
-              $newRecordColumnList
-            )
-        """.trimIndent()
+            |MERGE INTO ${fullyQualifiedName(targetTableName)} AS target_table
+            |USING (
+            |$selectSourceRecords
+            |) AS new_record
+            |ON $pkEquivalent
+            |$cdcDeleteClause
+            |WHEN MATCHED AND $cursorComparison THEN UPDATE SET
+            |  $columnAssignments
+            |WHEN NOT MATCHED $cdcSkipInsertClause THEN INSERT (
+            |  $columnList
+            |) VALUES (
+            |  $newRecordColumnList
+            |)
+        """.trimMargin()
             } else {
                 """
-            MERGE INTO ${snowflakeSqlNameUtils.fullyQualifiedName(targetTableName)} AS target_table
-            USING (
-              $selectSourceRecords
-            ) AS new_record
-            ON $pkEquivalent
-            WHEN MATCHED AND $cursorComparison THEN UPDATE SET
-              $columnAssignments
-            WHEN NOT MATCHED THEN INSERT (
-              $columnList
-            ) VALUES (
-              $newRecordColumnList
-            )
-        """.trimIndent()
+            |MERGE INTO ${fullyQualifiedName(targetTableName)} AS target_table
+            |USING (
+            |$selectSourceRecords
+            |) AS new_record
+            |ON $pkEquivalent
+            |WHEN MATCHED AND $cursorComparison THEN UPDATE SET
+            |  $columnAssignments
+            |WHEN NOT MATCHED THEN INSERT (
+            |  $columnList
+            |) VALUES (
+            |  $newRecordColumnList
+            |)
+        """.trimMargin()
             }
 
         return mergeStatement.andLog()
@@ -242,75 +246,66 @@ class SnowflakeDirectLoadSqlGenerator(
      * table. Uses ROW_NUMBER() window function to select the most recent record per primary key.
      */
     private fun selectDedupedRecords(
-        stream: DestinationStream,
-        sourceTableName: TableName,
-        columnNameMapping: ColumnNameMapping
+        tableSchema: StreamTableSchema,
+        sourceTableName: TableName
     ): String {
-        val columnList: String =
-            columnUtils
-                .getFormattedColumnNames(
-                    columns = stream.schema.asColumns(),
-                    columnNameMapping = columnNameMapping,
-                    quote = false,
-                )
-                .joinToString(
-                    ",\n",
-                ) {
-                    it.quote()
-                }
-        val importType = stream.importType as Dedupe
+        val allColumns = buildList {
+            add(SNOWFLAKE_AB_RAW_ID)
+            add(SNOWFLAKE_AB_EXTRACTED_AT)
+            add(SNOWFLAKE_AB_META)
+            add(SNOWFLAKE_AB_GENERATION_ID)
+            addAll(tableSchema.columnSchema.finalSchema.keys)
+        }
+        val columnList: String = allColumns.joinToString(",\n      ") { it.quote() }
 
         // Build the primary key list for partitioning
+        val pks = tableSchema.getPrimaryKey().flatten()
         val pkList =
-            if (importType.primaryKey.isNotEmpty()) {
-                importType.primaryKey.joinToString(",") { fieldPath ->
-                    (columnNameMapping[fieldPath.first()] ?: fieldPath.first()).quote()
-                }
+            if (pks.isNotEmpty()) {
+                pks.joinToString(",") { it.quote() }
             } else {
                 // Should not happen as we check this earlier, but handle it defensively
                 throw IllegalArgumentException("Cannot deduplicate without primary key")
             }
 
         // Build cursor order clause for sorting within each partition
+        val cursor = tableSchema.getCursor().firstOrNull()
         val cursorOrderClause =
-            if (importType.cursor.isNotEmpty()) {
-                val columnName =
-                    (columnNameMapping[importType.cursor.first()] ?: importType.cursor.first())
-                        .quote()
-                "$columnName DESC NULLS LAST,"
+            if (cursor != null) {
+                "${cursor.quote()} DESC NULLS LAST,"
             } else {
                 ""
             }
 
         return """
-            WITH records AS (
-              SELECT
-                $columnList
-              FROM ${snowflakeSqlNameUtils.fullyQualifiedName(sourceTableName)}
-            ), numbered_rows AS (
-              SELECT *, ROW_NUMBER() OVER (
-                PARTITION BY $pkList ORDER BY $cursorOrderClause "${COLUMN_NAME_AB_EXTRACTED_AT.toSnowflakeCompatibleName()}" DESC
-              ) AS row_number
-              FROM records
-            )
-            SELECT $columnList
-            FROM numbered_rows
-            WHERE row_number = 1
+            |  WITH records AS (
+            |    SELECT
+            |      $columnList
+            |    FROM ${fullyQualifiedName(sourceTableName)}
+            |  ), numbered_rows AS (
+            |    SELECT *, ROW_NUMBER() OVER (
+            |      PARTITION BY $pkList ORDER BY $cursorOrderClause "$SNOWFLAKE_AB_EXTRACTED_AT" DESC
+            |    ) AS row_number
+            |    FROM records
+            |  )
+            |  SELECT $columnList
+            |  FROM numbered_rows
+            |  WHERE row_number = 1
         """
-            .trimIndent()
+            .trimMargin()
             .andLog()
     }
 
     fun dropTable(tableName: TableName): String {
-        return "DROP TABLE IF EXISTS ${snowflakeSqlNameUtils.fullyQualifiedName(tableName)}".andLog()
+        return "DROP TABLE IF EXISTS ${fullyQualifiedName(tableName)}".andLog()
     }
 
     fun getGenerationId(
         tableName: TableName,
     ): String {
         return """
-            SELECT "${columnUtils.getGenerationIdColumnName()}"
-            FROM ${snowflakeSqlNameUtils.fullyQualifiedName(tableName)}
+            SELECT "${columnManager.getGenerationIdColumnName()}"
+            FROM ${fullyQualifiedName(tableName)}
             LIMIT 1
         """
             .trimIndent()
@@ -318,12 +313,12 @@ class SnowflakeDirectLoadSqlGenerator(
     }
 
     fun createSnowflakeStage(tableName: TableName): String {
-        val stageName = snowflakeSqlNameUtils.fullyQualifiedStageName(tableName)
+        val stageName = fullyQualifiedStageName(tableName)
         return "CREATE STAGE IF NOT EXISTS $stageName".andLog()
     }
 
     fun putInStage(tableName: TableName, tempFilePath: String): String {
-        val stageName = snowflakeSqlNameUtils.fullyQualifiedStageName(tableName, true)
+        val stageName = fullyQualifiedStageName(tableName, true)
         return """
             PUT 'file://$tempFilePath' '@$stageName'
             AUTO_COMPRESS = FALSE
@@ -334,35 +329,45 @@ class SnowflakeDirectLoadSqlGenerator(
             .andLog()
     }
 
-    fun copyFromStage(tableName: TableName, filename: String): String {
-        val stageName = snowflakeSqlNameUtils.fullyQualifiedStageName(tableName, true)
+    fun copyFromStage(
+        tableName: TableName,
+        filename: String,
+        columnNames: List<String>? = null
+    ): String {
+        val stageName = fullyQualifiedStageName(tableName, true)
+        val columnList =
+            columnNames?.let { names -> "(${names.joinToString(", ") { it.quote() }})" } ?: ""
 
         return """
-            COPY INTO ${snowflakeSqlNameUtils.fullyQualifiedName(tableName)}
-            FROM '@$stageName'
-            FILE_FORMAT = (
-                TYPE = 'CSV'
-                COMPRESSION = GZIP
-                FIELD_DELIMITER = '$CSV_FIELD_SEPARATOR'
-                RECORD_DELIMITER = '$CSV_LINE_DELIMITER'
-                FIELD_OPTIONALLY_ENCLOSED_BY = '"'
-                TRIM_SPACE = TRUE
-                ERROR_ON_COLUMN_COUNT_MISMATCH = FALSE
-                REPLACE_INVALID_CHARACTERS = TRUE
-                ESCAPE = NONE
-                ESCAPE_UNENCLOSED_FIELD = NONE
-            )
-            ON_ERROR = 'ABORT_STATEMENT'
-            PURGE = TRUE
-            files = ('$filename')
+            |COPY INTO ${fullyQualifiedName(tableName)}$columnList
+            |FROM '@$stageName'
+            |FILE_FORMAT = (
+            |    TYPE = 'CSV'
+            |    COMPRESSION = GZIP
+            |    FIELD_DELIMITER = '$CSV_FIELD_SEPARATOR'
+            |    RECORD_DELIMITER = '$CSV_LINE_DELIMITER'
+            |    FIELD_OPTIONALLY_ENCLOSED_BY = '"'
+            |    TRIM_SPACE = TRUE
+            |    ERROR_ON_COLUMN_COUNT_MISMATCH = FALSE
+            |    REPLACE_INVALID_CHARACTERS = TRUE
+            |    ESCAPE = NONE
+            |    ESCAPE_UNENCLOSED_FIELD = NONE
+            |)
+            |ON_ERROR = 'ABORT_STATEMENT'
+            |PURGE = TRUE
+            |files = ('$filename')
         """
-            .trimIndent()
+            .trimMargin()
             .andLog()
     }
 
     fun swapTableWith(sourceTableName: TableName, targetTableName: TableName): String {
         return """
-            ALTER TABLE ${snowflakeSqlNameUtils.fullyQualifiedName(sourceTableName)} SWAP WITH ${snowflakeSqlNameUtils.fullyQualifiedName(targetTableName)}
+            ALTER TABLE ${fullyQualifiedName(sourceTableName)} SWAP WITH ${
+            fullyQualifiedName(
+                targetTableName,
+            )
+        }
         """
             .trimIndent()
             .andLog()
@@ -372,7 +377,11 @@ class SnowflakeDirectLoadSqlGenerator(
         // Snowflake RENAME TO only accepts the table name, not a fully qualified name
         // The renamed table stays in the same schema
         return """
-            ALTER TABLE ${snowflakeSqlNameUtils.fullyQualifiedName(sourceTableName)} RENAME TO ${snowflakeSqlNameUtils.fullyQualifiedName(targetTableName)}
+            ALTER TABLE ${fullyQualifiedName(sourceTableName)} RENAME TO ${
+            fullyQualifiedName(
+                targetTableName,
+            )
+        }
         """
             .trimIndent()
             .andLog()
@@ -382,7 +391,7 @@ class SnowflakeDirectLoadSqlGenerator(
         schemaName: String,
         tableName: String,
     ): String =
-        """DESCRIBE TABLE ${snowflakeSqlNameUtils.fullyQualifiedName(TableName(schemaName, tableName))}""".andLog()
+        """DESCRIBE TABLE ${fullyQualifiedName(TableName(schemaName, tableName))}""".andLog()
 
     fun alterTable(
         tableName: TableName,
@@ -391,14 +400,14 @@ class SnowflakeDirectLoadSqlGenerator(
         modifiedColumns: Map<String, ColumnTypeChange>,
     ): Set<String> {
         val clauses = mutableSetOf<String>()
-        val prettyTableName = snowflakeSqlNameUtils.fullyQualifiedName(tableName)
+        val prettyTableName = fullyQualifiedName(tableName)
         addedColumns.forEach { (name, columnType) ->
             clauses.add(
                 // Note that we intentionally don't set NOT NULL.
                 // We're adding a new column, and we don't know what constitutes a reasonable
                 // default value for preexisting records.
                 // So we add the column as nullable.
-                "ALTER TABLE $prettyTableName ADD COLUMN ${name.quote()} ${columnType.type};".andLog()
+                "ALTER TABLE $prettyTableName ADD COLUMN ${name.quote()} ${columnType.type};".andLog(),
             )
         }
         deletedColumns.forEach {
@@ -412,35 +421,34 @@ class SnowflakeDirectLoadSqlGenerator(
                 val tempColumn = "${name}_${uuidGenerator.v4()}"
                 clauses.add(
                     // As above: we add the column as nullable.
-                    "ALTER TABLE $prettyTableName ADD COLUMN ${tempColumn.quote()} ${typeChange.newType.type};".andLog()
+                    "ALTER TABLE $prettyTableName ADD COLUMN ${tempColumn.quote()} ${typeChange.newType.type};".andLog(),
                 )
                 clauses.add(
-                    "UPDATE $prettyTableName SET ${tempColumn.quote()} = CAST(${name.quote()} AS ${typeChange.newType.type});".andLog()
+                    "UPDATE $prettyTableName SET ${tempColumn.quote()} = CAST(${name.quote()} AS ${typeChange.newType.type});".andLog(),
                 )
                 val backupColumn = "${tempColumn}_backup"
                 clauses.add(
                     """
                     ALTER TABLE $prettyTableName
                     RENAME COLUMN "$name" TO "$backupColumn";
-                    """.trimIndent()
+                    """.trimIndent(),
                 )
                 clauses.add(
                     """
                     ALTER TABLE $prettyTableName
                     RENAME COLUMN "$tempColumn" TO "$name";
-                    """.trimIndent()
+                    """.trimIndent(),
                 )
                 clauses.add(
-                    "ALTER TABLE $prettyTableName DROP COLUMN ${backupColumn.quote()};".andLog()
+                    "ALTER TABLE $prettyTableName DROP COLUMN ${backupColumn.quote()};".andLog(),
                 )
             } else if (!typeChange.originalType.nullable && typeChange.newType.nullable) {
                 // If the type is unchanged, we can change a column from NOT NULL to nullable.
                 // But we'll never do the reverse, because there's a decent chance that historical
-                // records
-                // had null values.
+                // records had null values.
                 // Users can always manually ALTER COLUMN ... SET NOT NULL if they want.
                 clauses.add(
-                    """ALTER TABLE $prettyTableName ALTER COLUMN "$name" DROP NOT NULL;""".andLog()
+                    """ALTER TABLE $prettyTableName ALTER COLUMN "$name" DROP NOT NULL;""".andLog(),
                 )
             } else {
                 log.info {
@@ -450,4 +458,45 @@ class SnowflakeDirectLoadSqlGenerator(
         }
         return clauses
     }
+
+    @VisibleForTesting
+    fun fullyQualifiedName(tableName: TableName): String =
+        combineParts(listOf(getDatabaseName(), tableName.namespace, tableName.name))
+
+    @VisibleForTesting
+    fun fullyQualifiedNamespace(namespace: String) =
+        combineParts(listOf(getDatabaseName(), namespace))
+
+    @VisibleForTesting
+    fun fullyQualifiedStageName(tableName: TableName, escape: Boolean = false): String {
+        val currentTableName =
+            if (escape) {
+                tableName.name
+            } else {
+                tableName.name
+            }
+        return combineParts(
+            parts =
+                listOf(
+                    getDatabaseName(),
+                    tableName.namespace,
+                    "$STAGE_NAME_PREFIX$currentTableName",
+                ),
+            escape = escape,
+        )
+    }
+
+    @VisibleForTesting
+    internal fun combineParts(parts: List<String>, escape: Boolean = false): String =
+        parts
+            .map { if (escape) sqlEscape(it) else it }
+            .joinToString(separator = ".") {
+                if (!it.startsWith(QUOTE)) {
+                    "$QUOTE$it$QUOTE"
+                } else {
+                    it
+                }
+            }
+
+    private fun getDatabaseName() = config.database.toSnowflakeCompatibleName()
 }

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/sql/SnowflakeSqlEscapeUtils.kt

@@ -0,0 +1,29 @@
+/*
+ * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
+ */
+
+package io.airbyte.integrations.destination.snowflake.sql
+
+const val STAGE_NAME_PREFIX = "airbyte_stage_"
+internal const val QUOTE: String = "\""
+
+fun sqlEscape(part: String) = part.replace("\\", "\\\\").replace("'", "\\'").replace("\"", "\\\"")
+
+/**
+ * Surrounds the string instance with double quotation marks (e.g. "some string" -> "\"some
+ * string\"").
+ */
+fun String.quote() = "$QUOTE$this$QUOTE"
+
+/**
+ * Escapes double-quotes in a JSON identifier by doubling them. This is legacy -- I don't know why
+ * this would be necessary but no harm in keeping it, so I am keeping it.
+ *
+ * @return The escaped identifier.
+ */
+fun escapeJsonIdentifier(identifier: String): String {
+    // Note that we don't need to escape backslashes here!
+    // The only special character in an identifier is the double-quote, which needs to be
+    // doubled.
+    return identifier.replace(QUOTE, "$QUOTE$QUOTE")
+}

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/sql/SnowflakeSqlNameUtils.kt

@@ -1,57 +0,0 @@
-/*
- * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
- */
-
-package io.airbyte.integrations.destination.snowflake.sql
-
-import io.airbyte.cdk.load.table.TableName
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
-import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
-import jakarta.inject.Singleton
-
-const val STAGE_NAME_PREFIX = "airbyte_stage_"
-internal const val QUOTE: String = "\""
-
-fun sqlEscape(part: String) = part.replace("\\", "\\\\").replace("'", "\\'").replace("\"", "\\\"")
-
-@Singleton
-class SnowflakeSqlNameUtils(
-    private val snowflakeConfiguration: SnowflakeConfiguration,
-) {
-    fun fullyQualifiedName(tableName: TableName): String =
-        combineParts(listOf(getDatabaseName(), tableName.namespace, tableName.name))
-
-    fun fullyQualifiedNamespace(namespace: String) =
-        combineParts(listOf(getDatabaseName(), namespace))
-
-    fun fullyQualifiedStageName(tableName: TableName, escape: Boolean = false): String {
-        val currentTableName =
-            if (escape) {
-                tableName.name
-            } else {
-                tableName.name
-            }
-        return combineParts(
-            parts =
-                listOf(
-                    getDatabaseName(),
-                    tableName.namespace,
-                    "$STAGE_NAME_PREFIX$currentTableName"
-                ),
-            escape = escape,
-        )
-    }
-
-    fun combineParts(parts: List<String>, escape: Boolean = false): String =
-        parts
-            .map { if (escape) sqlEscape(it) else it }
-            .joinToString(separator = ".") {
-                if (!it.startsWith(QUOTE)) {
-                    "$QUOTE$it$QUOTE"
-                } else {
-                    it
-                }
-            }
-
-    private fun getDatabaseName() = snowflakeConfiguration.database.toSnowflakeCompatibleName()
-}

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/write/SnowflakeWriter.kt

@@ -6,38 +6,39 @@ package io.airbyte.integrations.destination.snowflake.write
 
 import io.airbyte.cdk.SystemErrorException
 import io.airbyte.cdk.load.command.Dedupe
+import io.airbyte.cdk.load.command.DestinationCatalog
 import io.airbyte.cdk.load.command.DestinationStream
-import io.airbyte.cdk.load.orchestration.db.DatabaseInitialStatusGatherer
-import io.airbyte.cdk.load.orchestration.db.TempTableNameGenerator
-import io.airbyte.cdk.load.orchestration.db.direct_load_table.DirectLoadInitialStatus
-import io.airbyte.cdk.load.orchestration.db.direct_load_table.DirectLoadTableAppendStreamLoader
-import io.airbyte.cdk.load.orchestration.db.direct_load_table.DirectLoadTableAppendTruncateStreamLoader
-import io.airbyte.cdk.load.orchestration.db.direct_load_table.DirectLoadTableDedupStreamLoader
-import io.airbyte.cdk.load.orchestration.db.direct_load_table.DirectLoadTableDedupTruncateStreamLoader
-import io.airbyte.cdk.load.orchestration.db.direct_load_table.DirectLoadTableExecutionConfig
-import io.airbyte.cdk.load.orchestration.db.legacy_typing_deduping.TableCatalog
+import io.airbyte.cdk.load.table.ColumnNameMapping
+import io.airbyte.cdk.load.table.DatabaseInitialStatusGatherer
+import io.airbyte.cdk.load.table.TempTableNameGenerator
+import io.airbyte.cdk.load.table.directload.DirectLoadInitialStatus
+import io.airbyte.cdk.load.table.directload.DirectLoadTableAppendStreamLoader
+import io.airbyte.cdk.load.table.directload.DirectLoadTableAppendTruncateStreamLoader
+import io.airbyte.cdk.load.table.directload.DirectLoadTableDedupStreamLoader
+import io.airbyte.cdk.load.table.directload.DirectLoadTableDedupTruncateStreamLoader
+import io.airbyte.cdk.load.table.directload.DirectLoadTableExecutionConfig
 import io.airbyte.cdk.load.write.DestinationWriter
 import io.airbyte.cdk.load.write.StreamLoader
 import io.airbyte.cdk.load.write.StreamStateStore
 import io.airbyte.integrations.destination.snowflake.client.SnowflakeAirbyteClient
-import io.airbyte.integrations.destination.snowflake.db.escapeJsonIdentifier
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
+import io.airbyte.integrations.destination.snowflake.sql.escapeJsonIdentifier
 import jakarta.inject.Singleton
 
 @Singleton
 class SnowflakeWriter(
-    private val names: TableCatalog,
+    private val catalog: DestinationCatalog,
     private val stateGatherer: DatabaseInitialStatusGatherer<DirectLoadInitialStatus>,
     private val streamStateStore: StreamStateStore<DirectLoadTableExecutionConfig>,
     private val snowflakeClient: SnowflakeAirbyteClient,
-    private val tempTableNameGenerator: TempTableNameGenerator,
     private val snowflakeConfiguration: SnowflakeConfiguration,
+    private val tempTableNameGenerator: TempTableNameGenerator,
 ) : DestinationWriter {
     private lateinit var initialStatuses: Map<DestinationStream, DirectLoadInitialStatus>
 
     override suspend fun setup() {
-        names.values
-            .map { (tableNames, _) -> tableNames.finalTableName!!.namespace }
+        catalog.streams
+            .map { it.tableSchema.tableNames.finalTableName!!.namespace }
             .toSet()
             .forEach { snowflakeClient.createNamespace(it) }
 
@@ -45,15 +46,15 @@ class SnowflakeWriter(
             escapeJsonIdentifier(snowflakeConfiguration.internalTableSchema)
         )
 
-        initialStatuses = stateGatherer.gatherInitialStatus(names)
+        initialStatuses = stateGatherer.gatherInitialStatus()
     }
 
     override fun createStreamLoader(stream: DestinationStream): StreamLoader {
         val initialStatus = initialStatuses[stream]!!
-        val tableNameInfo = names[stream]!!
-        val realTableName = tableNameInfo.tableNames.finalTableName!!
-        val tempTableName = tempTableNameGenerator.generate(realTableName)
-        val columnNameMapping = tableNameInfo.columnNameMapping
+        val realTableName = stream.tableSchema.tableNames.finalTableName!!
+        val tempTableName = stream.tableSchema.tableNames.tempTableName!!
+        val columnNameMapping =
+            ColumnNameMapping(stream.tableSchema.columnSchema.inputToFinalColumnNames)
         return when (stream.minimumGenerationId) {
             0L ->
                 when (stream.importType) {

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/write/load/SnowflakeInsertBuffer.kt

@@ -9,11 +9,12 @@ import de.siegmar.fastcsv.writer.CsvWriter
 import de.siegmar.fastcsv.writer.LineDelimiter
 import de.siegmar.fastcsv.writer.QuoteStrategies
 import io.airbyte.cdk.load.data.AirbyteValue
-import io.airbyte.cdk.load.table.TableName
+import io.airbyte.cdk.load.schema.model.ColumnSchema
+import io.airbyte.cdk.load.schema.model.TableName
 import io.airbyte.integrations.destination.snowflake.client.SnowflakeAirbyteClient
+import io.airbyte.integrations.destination.snowflake.schema.SnowflakeColumnManager
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
 import io.airbyte.integrations.destination.snowflake.sql.QUOTE
-import io.airbyte.integrations.destination.snowflake.sql.SnowflakeColumnUtils
 import io.github.oshai.kotlinlogging.KotlinLogging
 import java.io.File
 import java.io.OutputStream
@@ -36,10 +37,11 @@ private const val CSV_WRITER_BUFFER_SIZE = 1024 * 1024 // 1 MB
 
 class SnowflakeInsertBuffer(
     private val tableName: TableName,
-    val columns: LinkedHashMap<String, String>,
     private val snowflakeClient: SnowflakeAirbyteClient,
     val snowflakeConfiguration: SnowflakeConfiguration,
-    val snowflakeColumnUtils: SnowflakeColumnUtils,
+    val columnSchema: ColumnSchema,
+    private val columnManager: SnowflakeColumnManager,
+    private val snowflakeRecordFormatter: SnowflakeRecordFormatter,
     private val flushLimit: Int = DEFAULT_FLUSH_LIMIT,
 ) {
 
@@ -57,12 +59,6 @@ class SnowflakeInsertBuffer(
             .lineDelimiter(CSV_LINE_DELIMITER)
             .quoteStrategy(QuoteStrategies.REQUIRED)
 
-    private val snowflakeRecordFormatter: SnowflakeRecordFormatter =
-        when (snowflakeConfiguration.legacyRawTablesOnly) {
-            true -> SnowflakeRawRecordFormatter(columns, snowflakeColumnUtils)
-            else -> SnowflakeSchemaRecordFormatter(columns, snowflakeColumnUtils)
-        }
-
     fun accumulate(recordFields: Map<String, AirbyteValue>) {
         if (csvFilePath == null) {
             val csvFile = createCsvFile()
@@ -92,7 +88,9 @@ class SnowflakeInsertBuffer(
                     "Copying staging data into ${tableName.toPrettyString(quote = QUOTE)}..."
                 }
                 // Finally, copy the data from the staging table to the final table
-                snowflakeClient.copyFromStage(tableName, filePath.fileName.toString())
+                // Pass column names to ensure correct mapping even after ALTER TABLE operations
+                val columnNames = columnManager.getTableColumnNames(columnSchema)
+                snowflakeClient.copyFromStage(tableName, filePath.fileName.toString(), columnNames)
                 logger.info {
                     "Finished insert of $recordCount row(s) into ${tableName.toPrettyString(quote = QUOTE)}."
                 }
@@ -117,7 +115,9 @@ class SnowflakeInsertBuffer(
 
     private fun writeToCsvFile(record: Map<String, AirbyteValue>) {
         csvWriter?.let {
-            it.writeRecord(snowflakeRecordFormatter.format(record).map { col -> col.toString() })
+            it.writeRecord(
+                snowflakeRecordFormatter.format(record, columnSchema).map { col -> col.toString() }
+            )
             recordCount++
             if ((recordCount % flushLimit) == 0) {
                 it.flush()

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/write/load/SnowflakeRecordFormatter.kt

@@ -8,99 +8,62 @@ import io.airbyte.cdk.load.data.AirbyteValue
 import io.airbyte.cdk.load.data.NullValue
 import io.airbyte.cdk.load.data.StringValue
 import io.airbyte.cdk.load.data.csv.toCsvValue
-import io.airbyte.cdk.load.message.Meta
+import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_EXTRACTED_AT
+import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_GENERATION_ID
+import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_LOADED_AT
+import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_META
+import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_RAW_ID
+import io.airbyte.cdk.load.schema.model.ColumnSchema
 import io.airbyte.cdk.load.util.Jsons
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
-import io.airbyte.integrations.destination.snowflake.sql.SnowflakeColumnUtils
 
 interface SnowflakeRecordFormatter {
-    fun format(record: Map<String, AirbyteValue>): List<Any>
+    fun format(record: Map<String, AirbyteValue>, columnSchema: ColumnSchema): List<Any>
 }
 
-class SnowflakeSchemaRecordFormatter(
-    private val columns: LinkedHashMap<String, String>,
-    val snowflakeColumnUtils: SnowflakeColumnUtils,
-) : SnowflakeRecordFormatter {
+class SnowflakeSchemaRecordFormatter : SnowflakeRecordFormatter {
+    override fun format(record: Map<String, AirbyteValue>, columnSchema: ColumnSchema): List<Any> {
+        val result = mutableListOf<Any>()
+        val userColumns = columnSchema.finalSchema.keys
 
-    private val airbyteColumnNames =
-        snowflakeColumnUtils.getFormattedDefaultColumnNames(false).toSet()
+        // WARNING:  MUST match the order defined in SnowflakeColumnManager#getTableColumnNames
+        //
+        // Why don't we just use that here? Well, unlike the user fields, the meta fields on the
+        // record are not munged for the destination. So we must access the values for those columns
+        // using the original lowercase meta key.
+        result.add(record[COLUMN_NAME_AB_RAW_ID].toCsvValue())
+        result.add(record[COLUMN_NAME_AB_EXTRACTED_AT].toCsvValue())
+        result.add(record[COLUMN_NAME_AB_META].toCsvValue())
+        result.add(record[COLUMN_NAME_AB_GENERATION_ID].toCsvValue())
 
-    override fun format(record: Map<String, AirbyteValue>): List<Any> =
-        columns.map { (columnName, _) ->
-            /*
-             * Meta columns are forced to uppercase for backwards compatibility with previous
-             * versions of the destination.  Therefore, convert the column to lowercase so
-             * that it can match the constants, which use the lowercase version of the meta
-             * column names.
-             */
-            if (airbyteColumnNames.contains(columnName)) {
-                record[columnName.lowercase()].toCsvValue()
-            } else {
-                record.keys
-                    // The columns retrieved from Snowflake do not have any escaping applied.
-                    // Therefore, re-apply the compatible name escaping to the name of the
-                    // columns retrieved from Snowflake.  The record keys should already have
-                    // been escaped by the CDK before arriving at the aggregate, so no need
-                    // to escape again here.
-                    .find { it == columnName.toSnowflakeCompatibleName() }
-                    ?.let { record[it].toCsvValue() }
-                    ?: ""
-            }
-        }
+        // Add user columns from the final schema
+        userColumns.forEach { columnName -> result.add(record[columnName].toCsvValue()) }
+
+        return result
+    }
 }
 
-class SnowflakeRawRecordFormatter(
-    columns: LinkedHashMap<String, String>,
-    val snowflakeColumnUtils: SnowflakeColumnUtils,
-) : SnowflakeRecordFormatter {
-    private val columns = columns.keys
+class SnowflakeRawRecordFormatter : SnowflakeRecordFormatter {
 
-    private val airbyteColumnNames =
-        snowflakeColumnUtils.getFormattedDefaultColumnNames(false).toSet()
-
-    override fun format(record: Map<String, AirbyteValue>): List<Any> =
+    override fun format(record: Map<String, AirbyteValue>, columnSchema: ColumnSchema): List<Any> =
         toOutputRecord(record.toMutableMap())
 
     private fun toOutputRecord(record: MutableMap<String, AirbyteValue>): List<Any> {
         val outputRecord = mutableListOf<Any>()
-        // Copy the Airbyte metadata columns to the raw output, removing each
-        // one from the record to avoid duplicates in the "data" field
-        columns
-            .filter { airbyteColumnNames.contains(it) && it != Meta.COLUMN_NAME_DATA }
-            .forEach { column -> safeAddToOutput(column, record, outputRecord) }
+        val mutableRecord = record.toMutableMap()
+
+        // Add meta columns in order (except _airbyte_data which we handle specially)
+        outputRecord.add(mutableRecord.remove(COLUMN_NAME_AB_RAW_ID)?.toCsvValue() ?: "")
+        outputRecord.add(mutableRecord.remove(COLUMN_NAME_AB_EXTRACTED_AT)?.toCsvValue() ?: "")
+        outputRecord.add(mutableRecord.remove(COLUMN_NAME_AB_META)?.toCsvValue() ?: "")
+        outputRecord.add(mutableRecord.remove(COLUMN_NAME_AB_GENERATION_ID)?.toCsvValue() ?: "")
+        outputRecord.add(mutableRecord.remove(COLUMN_NAME_AB_LOADED_AT)?.toCsvValue() ?: "")
+
         // Do not output null values in the JSON raw output
-        val filteredRecord = record.filter { (_, v) -> v !is NullValue }
-        // Convert all the remaining columns in the record to a JSON document stored in the "data"
-        // column.  Add it in the same position as the _airbyte_data column in the column list to
-        // ensure it is inserted into the proper column in the table.
-        insert(
-            columns.indexOf(Meta.COLUMN_NAME_DATA),
-            StringValue(Jsons.writeValueAsString(filteredRecord)).toCsvValue(),
-            outputRecord
-        )
+        val filteredRecord = mutableRecord.filter { (_, v) -> v !is NullValue }
+
+        // Convert all the remaining columns to a JSON document stored in the "data" column
+        outputRecord.add(StringValue(Jsons.writeValueAsString(filteredRecord)).toCsvValue())
+
         return outputRecord
     }
-
-    private fun safeAddToOutput(
-        key: String,
-        record: MutableMap<String, AirbyteValue>,
-        output: MutableList<Any>
-    ) {
-        val extractedValue = record.remove(key)
-        // Ensure that the data is inserted into the list at the same position as the column
-        insert(columns.indexOf(key), extractedValue?.toCsvValue() ?: "", output)
-    }
-
-    private fun insert(index: Int, value: Any, list: MutableList<Any>) {
-        /*
-         * Attempt to insert the value into the proper order in the list.  If the index
-         * is already present in the list, use the add(index, element) method to insert it
-         * into the proper order and push everything to the right.  If the index is at the
-         * end of the list, just use add(element) to insert it at the end.  If the index
-         * is further beyond the end of the list, throw an exception as that should not occur.
-         */
-        if (index < list.size) list.add(index, value)
-        else if (index == list.size || index == list.size + 1) list.add(value)
-        else throw IndexOutOfBoundsException()
-    }
 }

airbyte-integrations/connectors/destination-snowflake/src/main/kotlin/io/airbyte/integrations/destination/snowflake/write/transform/SnowflakeColumnNameMapper.kt

@@ -1,25 +0,0 @@
-/*
- * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
- */
-
-package io.airbyte.integrations.destination.snowflake.write.transform
-
-import io.airbyte.cdk.load.command.DestinationStream
-import io.airbyte.cdk.load.dataflow.transform.ColumnNameMapper
-import io.airbyte.cdk.load.orchestration.db.legacy_typing_deduping.TableCatalog
-import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
-import jakarta.inject.Singleton
-
-@Singleton
-class SnowflakeColumnNameMapper(
-    private val catalogInfo: TableCatalog,
-    private val snowflakeConfiguration: SnowflakeConfiguration,
-) : ColumnNameMapper {
-    override fun getMappedColumnName(stream: DestinationStream, columnName: String): String {
-        if (snowflakeConfiguration.legacyRawTablesOnly == true) {
-            return columnName
-        } else {
-            return catalogInfo.getMappedColumnName(stream, columnName)!!
-        }
-    }
-}

airbyte-integrations/connectors/destination-snowflake/src/test-integration/kotlin/io/airbyte/integrations/destination/snowflake/component/SnowflakeTableOperationsTest.kt

@@ -7,9 +7,11 @@ package io.airbyte.integrations.destination.snowflake.component
 import io.airbyte.cdk.load.component.TableOperationsFixtures
 import io.airbyte.cdk.load.component.TableOperationsSuite
 import io.airbyte.cdk.load.message.Meta
+import io.airbyte.cdk.load.schema.TableSchemaFactory
 import io.airbyte.integrations.destination.snowflake.client.SnowflakeAirbyteClient
-import io.airbyte.integrations.destination.snowflake.component.SnowflakeComponentTestFixtures.idTestWithCdcMapping
-import io.airbyte.integrations.destination.snowflake.component.SnowflakeComponentTestFixtures.testMapping
+import io.airbyte.integrations.destination.snowflake.component.config.SnowflakeComponentTestFixtures.idTestWithCdcMapping
+import io.airbyte.integrations.destination.snowflake.component.config.SnowflakeComponentTestFixtures.testMapping
+import io.airbyte.integrations.destination.snowflake.component.config.SnowflakeTestTableOperationsClient
 import io.micronaut.test.extensions.junit5.annotation.MicronautTest
 import org.junit.jupiter.api.Test
 import org.junit.jupiter.api.parallel.Execution
@@ -20,6 +22,7 @@ import org.junit.jupiter.api.parallel.ExecutionMode
 class SnowflakeTableOperationsTest(
     override val client: SnowflakeAirbyteClient,
     override val testClient: SnowflakeTestTableOperationsClient,
+    override val schemaFactory: TableSchemaFactory,
 ) : TableOperationsSuite {
     override val airbyteMetaColumnMapping = Meta.COLUMN_NAMES.associateWith { it.uppercase() }
 

airbyte-integrations/connectors/destination-snowflake/src/test-integration/kotlin/io/airbyte/integrations/destination/snowflake/component/SnowflakeTableSchemaEvolutionTest.kt

@@ -9,13 +9,15 @@ import io.airbyte.cdk.load.component.TableSchemaEvolutionFixtures
 import io.airbyte.cdk.load.component.TableSchemaEvolutionSuite
 import io.airbyte.cdk.load.data.StringValue
 import io.airbyte.cdk.load.message.Meta
+import io.airbyte.cdk.load.schema.TableSchemaFactory
 import io.airbyte.cdk.load.table.ColumnNameMapping
 import io.airbyte.cdk.load.util.serializeToString
 import io.airbyte.integrations.destination.snowflake.client.SnowflakeAirbyteClient
-import io.airbyte.integrations.destination.snowflake.component.SnowflakeComponentTestFixtures.allTypesColumnNameMapping
-import io.airbyte.integrations.destination.snowflake.component.SnowflakeComponentTestFixtures.allTypesTableSchema
-import io.airbyte.integrations.destination.snowflake.component.SnowflakeComponentTestFixtures.idAndTestMapping
-import io.airbyte.integrations.destination.snowflake.component.SnowflakeComponentTestFixtures.testMapping
+import io.airbyte.integrations.destination.snowflake.component.config.SnowflakeComponentTestFixtures.allTypesColumnNameMapping
+import io.airbyte.integrations.destination.snowflake.component.config.SnowflakeComponentTestFixtures.allTypesTableSchema
+import io.airbyte.integrations.destination.snowflake.component.config.SnowflakeComponentTestFixtures.idAndTestMapping
+import io.airbyte.integrations.destination.snowflake.component.config.SnowflakeComponentTestFixtures.testMapping
+import io.airbyte.integrations.destination.snowflake.component.config.SnowflakeTestTableOperationsClient
 import io.micronaut.test.extensions.junit5.annotation.MicronautTest
 import org.junit.jupiter.api.Test
 import org.junit.jupiter.api.parallel.Execution
@@ -27,6 +29,7 @@ class SnowflakeTableSchemaEvolutionTest(
     override val client: SnowflakeAirbyteClient,
     override val opsClient: SnowflakeAirbyteClient,
     override val testClient: SnowflakeTestTableOperationsClient,
+    override val schemaFactory: TableSchemaFactory,
 ) : TableSchemaEvolutionSuite {
     override val airbyteMetaColumnMapping = Meta.COLUMN_NAMES.associateWith { it.uppercase() }
 

airbyte-integrations/connectors/destination-snowflake/src/test-integration/kotlin/io/airbyte/integrations/destination/snowflake/component/config/SnowflakeComponentTestFixtures.kt

@@ -2,7 +2,7 @@
  * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
  */
 
-package io.airbyte.integrations.destination.snowflake.component
+package io.airbyte.integrations.destination.snowflake.component.config
 
 import io.airbyte.cdk.load.component.ColumnType
 import io.airbyte.cdk.load.component.TableOperationsFixtures
@@ -33,6 +33,8 @@ object SnowflakeComponentTestFixtures {
                 "TIME_NTZ" to ColumnType("TIME", true),
                 "ARRAY" to ColumnType("ARRAY", true),
                 "OBJECT" to ColumnType("OBJECT", true),
+                "UNION" to ColumnType("VARIANT", true),
+                "LEGACY_UNION" to ColumnType("VARIANT", true),
                 "UNKNOWN" to ColumnType("VARIANT", true),
             )
         )

airbyte-integrations/connectors/destination-snowflake/src/test-integration/kotlin/io/airbyte/integrations/destination/snowflake/component/config/SnowflakeTestConfigFactory.kt

@@ -2,7 +2,7 @@
  * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
  */
 
-package io.airbyte.integrations.destination.snowflake.component
+package io.airbyte.integrations.destination.snowflake.component.config
 
 import io.airbyte.cdk.load.component.config.TestConfigLoader.loadTestConfig
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration

airbyte-integrations/connectors/destination-snowflake/src/test-integration/kotlin/io/airbyte/integrations/destination/snowflake/component/config/SnowflakeTestTableOperationsClient.kt

@@ -2,22 +2,24 @@
  * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
  */
 
-package io.airbyte.integrations.destination.snowflake.component
+package io.airbyte.integrations.destination.snowflake.component.config
 
+import io.airbyte.cdk.load.component.ColumnType
 import io.airbyte.cdk.load.component.TestTableOperationsClient
 import io.airbyte.cdk.load.data.AirbyteValue
 import io.airbyte.cdk.load.dataflow.state.PartitionKey
 import io.airbyte.cdk.load.dataflow.transform.RecordDTO
-import io.airbyte.cdk.load.table.TableName
+import io.airbyte.cdk.load.schema.model.TableName
 import io.airbyte.cdk.load.util.Jsons
 import io.airbyte.integrations.destination.snowflake.client.SnowflakeAirbyteClient
 import io.airbyte.integrations.destination.snowflake.client.execute
 import io.airbyte.integrations.destination.snowflake.dataflow.SnowflakeAggregate
+import io.airbyte.integrations.destination.snowflake.schema.SnowflakeColumnManager
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
-import io.airbyte.integrations.destination.snowflake.sql.SnowflakeColumnUtils
-import io.airbyte.integrations.destination.snowflake.sql.SnowflakeSqlNameUtils
+import io.airbyte.integrations.destination.snowflake.sql.SnowflakeDirectLoadSqlGenerator
 import io.airbyte.integrations.destination.snowflake.sql.andLog
 import io.airbyte.integrations.destination.snowflake.write.load.SnowflakeInsertBuffer
+import io.airbyte.integrations.destination.snowflake.write.load.SnowflakeRecordFormatter
 import io.micronaut.context.annotation.Requires
 import jakarta.inject.Singleton
 import java.time.format.DateTimeFormatter
@@ -29,25 +31,40 @@ import net.snowflake.client.jdbc.SnowflakeTimestampWithTimezone
 class SnowflakeTestTableOperationsClient(
     private val client: SnowflakeAirbyteClient,
     private val dataSource: DataSource,
-    private val snowflakeSqlNameUtils: SnowflakeSqlNameUtils,
-    private val snowflakeColumnUtils: SnowflakeColumnUtils,
+    private val sqlGenerator: SnowflakeDirectLoadSqlGenerator,
     private val snowflakeConfiguration: SnowflakeConfiguration,
+    private val columnManager: SnowflakeColumnManager,
+    private val snowflakeRecordFormatter: SnowflakeRecordFormatter,
 ) : TestTableOperationsClient {
     override suspend fun dropNamespace(namespace: String) {
         dataSource.execute(
-            "DROP SCHEMA IF EXISTS ${snowflakeSqlNameUtils.fullyQualifiedNamespace(namespace)}".andLog()
+            "DROP SCHEMA IF EXISTS ${sqlGenerator.fullyQualifiedNamespace(namespace)}".andLog()
         )
     }
 
     override suspend fun insertRecords(table: TableName, records: List<Map<String, AirbyteValue>>) {
+        // TODO: we should just pass a proper column schema
+        // Since we don't pass in a proper column schema, we have to recreate one here
+        // Fetch the columns and filter out the meta columns so we're just looking at user columns
+        val columnTypes =
+            client.describeTable(table).filterNot {
+                columnManager.getMetaColumnNames().contains(it.key)
+            }
+        val columnSchema =
+            io.airbyte.cdk.load.schema.model.ColumnSchema(
+                inputToFinalColumnNames = columnTypes.keys.associateWith { it },
+                finalSchema = columnTypes.mapValues { (_, _) -> ColumnType("", true) },
+                inputSchema = emptyMap() // Not needed for insert buffer
+            )
         val a =
             SnowflakeAggregate(
                 SnowflakeInsertBuffer(
-                    table,
-                    client.describeTable(table),
-                    client,
-                    snowflakeConfiguration,
-                    snowflakeColumnUtils,
+                    tableName = table,
+                    snowflakeClient = client,
+                    snowflakeConfiguration = snowflakeConfiguration,
+                    columnSchema = columnSchema,
+                    columnManager = columnManager,
+                    snowflakeRecordFormatter = snowflakeRecordFormatter,
                 )
             )
         records.forEach { a.accept(RecordDTO(it, PartitionKey(""), 0, 0)) }

airbyte-integrations/connectors/destination-snowflake/src/test-integration/kotlin/io/airbyte/integrations/destination/snowflake/write/SnowflakeDataCleaner.kt

@@ -7,11 +7,11 @@ package io.airbyte.integrations.destination.snowflake.write
 import io.airbyte.cdk.load.test.util.DestinationCleaner
 import io.airbyte.integrations.destination.snowflake.SnowflakeBeanFactory
 import io.airbyte.integrations.destination.snowflake.cdk.SnowflakeMigratingConfigurationSpecificationSupplier
-import io.airbyte.integrations.destination.snowflake.db.escapeJsonIdentifier
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
+import io.airbyte.integrations.destination.snowflake.schema.toSnowflakeCompatibleName
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfigurationFactory
 import io.airbyte.integrations.destination.snowflake.sql.STAGE_NAME_PREFIX
+import io.airbyte.integrations.destination.snowflake.sql.escapeJsonIdentifier
 import io.airbyte.integrations.destination.snowflake.sql.quote
 import java.nio.file.Files
 import java.sql.Connection

airbyte-integrations/connectors/destination-snowflake/src/test-integration/kotlin/io/airbyte/integrations/destination/snowflake/write/SnowflakeDataDumper.kt

@@ -12,16 +12,22 @@ import io.airbyte.cdk.load.data.ObjectValue
 import io.airbyte.cdk.load.data.json.toAirbyteValue
 import io.airbyte.cdk.load.message.Meta
 import io.airbyte.cdk.load.table.CDC_DELETED_AT_COLUMN
+import io.airbyte.cdk.load.table.DefaultTempTableNameGenerator
 import io.airbyte.cdk.load.test.util.DestinationDataDumper
 import io.airbyte.cdk.load.test.util.OutputRecord
+import io.airbyte.cdk.load.util.UUIDGenerator
 import io.airbyte.cdk.load.util.deserializeToNode
 import io.airbyte.integrations.destination.snowflake.SnowflakeBeanFactory
-import io.airbyte.integrations.destination.snowflake.db.SnowflakeFinalTableNameGenerator
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
+import io.airbyte.integrations.destination.snowflake.schema.SnowflakeColumnManager
+import io.airbyte.integrations.destination.snowflake.schema.SnowflakeTableSchemaMapper
+import io.airbyte.integrations.destination.snowflake.schema.toSnowflakeCompatibleName
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
-import io.airbyte.integrations.destination.snowflake.sql.SnowflakeSqlNameUtils
+import io.airbyte.integrations.destination.snowflake.sql.SnowflakeDirectLoadSqlGenerator
 import io.airbyte.integrations.destination.snowflake.sql.sqlEscape
 import java.math.BigDecimal
+import java.sql.Date
+import java.sql.Time
+import java.sql.Timestamp
 import net.snowflake.client.jdbc.SnowflakeTimestampWithTimezone
 
 private val AIRBYTE_META_COLUMNS = Meta.COLUMN_NAMES + setOf(CDC_DELETED_AT_COLUMN)
@@ -34,8 +40,14 @@ class SnowflakeDataDumper(
         stream: DestinationStream
     ): List<OutputRecord> {
         val config = configProvider(spec)
-        val sqlUtils = SnowflakeSqlNameUtils(config)
-        val snowflakeFinalTableNameGenerator = SnowflakeFinalTableNameGenerator(config)
+        val snowflakeFinalTableNameGenerator =
+            SnowflakeTableSchemaMapper(
+                config = config,
+                tempTableNameGenerator = DefaultTempTableNameGenerator(),
+            )
+        val snowflakeColumnManager = SnowflakeColumnManager(config)
+        val sqlGenerator =
+            SnowflakeDirectLoadSqlGenerator(UUIDGenerator(), config, snowflakeColumnManager)
         val dataSource =
             SnowflakeBeanFactory()
                 .snowflakeDataSource(snowflakeConfiguration = config, airbyteEdition = "COMMUNITY")
@@ -46,7 +58,7 @@ class SnowflakeDataDumper(
             ds.connection.use { connection ->
                 val statement = connection.createStatement()
                 val tableName =
-                    snowflakeFinalTableNameGenerator.getTableName(stream.mappedDescriptor)
+                    snowflakeFinalTableNameGenerator.toFinalTableName(stream.mappedDescriptor)
 
                 // First check if the table exists
                 val tableExistsQuery =
@@ -69,7 +81,7 @@ class SnowflakeDataDumper(
 
                 val resultSet =
                     statement.executeQuery(
-                        "SELECT * FROM ${sqlUtils.fullyQualifiedName(tableName)}"
+                        "SELECT * FROM ${sqlGenerator.fullyQualifiedName(tableName)}"
                     )
 
                 while (resultSet.next()) {
@@ -143,10 +155,10 @@ class SnowflakeDataDumper(
     private fun convertValue(value: Any?): Any? =
         when (value) {
             is BigDecimal -> value.toBigInteger()
-            is java.sql.Date -> value.toLocalDate()
+            is Date -> value.toLocalDate()
             is SnowflakeTimestampWithTimezone -> value.toZonedDateTime()
-            is java.sql.Time -> value.toLocalTime()
-            is java.sql.Timestamp -> value.toLocalDateTime()
+            is Time -> value.toLocalTime()
+            is Timestamp -> value.toLocalDateTime()
             else -> value
         }
 }

airbyte-integrations/connectors/destination-snowflake/src/test-integration/kotlin/io/airbyte/integrations/destination/snowflake/write/SnowflakeExpectedRecordMapper.kt

@@ -15,7 +15,7 @@ import io.airbyte.cdk.load.dataflow.transform.ValidationResult
 import io.airbyte.cdk.load.message.Meta
 import io.airbyte.cdk.load.test.util.ExpectedRecordMapper
 import io.airbyte.cdk.load.test.util.OutputRecord
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
+import io.airbyte.integrations.destination.snowflake.schema.toSnowflakeCompatibleName
 import io.airbyte.integrations.destination.snowflake.write.transform.SnowflakeValueCoercer
 import io.airbyte.protocol.models.v0.AirbyteRecordMessageMetaChange
 import io.airbyte.protocol.models.v0.AirbyteRecordMessageMetaChange.Change

airbyte-integrations/connectors/destination-snowflake/src/test-integration/kotlin/io/airbyte/integrations/destination/snowflake/write/SnowflakeNameMapper.kt

@@ -5,7 +5,7 @@
 package io.airbyte.integrations.destination.snowflake.write
 
 import io.airbyte.cdk.load.test.util.NameMapper
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
+import io.airbyte.integrations.destination.snowflake.schema.toSnowflakeCompatibleName
 
 class SnowflakeNameMapper : NameMapper {
     override fun mapFieldName(path: List<String>): List<String> =

airbyte-integrations/connectors/destination-snowflake/src/test-integration/kotlin/io/airbyte/integrations/destination/snowflake/write/SnowflakeRawDataDumper.kt

@@ -9,13 +9,16 @@ import io.airbyte.cdk.load.command.DestinationStream
 import io.airbyte.cdk.load.data.StringValue
 import io.airbyte.cdk.load.data.json.toAirbyteValue
 import io.airbyte.cdk.load.message.Meta
+import io.airbyte.cdk.load.table.DefaultTempTableNameGenerator
 import io.airbyte.cdk.load.test.util.DestinationDataDumper
 import io.airbyte.cdk.load.test.util.OutputRecord
+import io.airbyte.cdk.load.util.UUIDGenerator
 import io.airbyte.cdk.load.util.deserializeToNode
 import io.airbyte.integrations.destination.snowflake.SnowflakeBeanFactory
-import io.airbyte.integrations.destination.snowflake.db.SnowflakeFinalTableNameGenerator
+import io.airbyte.integrations.destination.snowflake.schema.SnowflakeColumnManager
+import io.airbyte.integrations.destination.snowflake.schema.SnowflakeTableSchemaMapper
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
-import io.airbyte.integrations.destination.snowflake.sql.SnowflakeSqlNameUtils
+import io.airbyte.integrations.destination.snowflake.sql.SnowflakeDirectLoadSqlGenerator
 
 class SnowflakeRawDataDumper(
     private val configProvider: (ConfigurationSpecification) -> SnowflakeConfiguration
@@ -27,8 +30,18 @@ class SnowflakeRawDataDumper(
         val output = mutableListOf<OutputRecord>()
 
         val config = configProvider(spec)
-        val sqlUtils = SnowflakeSqlNameUtils(config)
-        val snowflakeFinalTableNameGenerator = SnowflakeFinalTableNameGenerator(config)
+        val snowflakeColumnManager = SnowflakeColumnManager(config)
+        val sqlGenerator =
+            SnowflakeDirectLoadSqlGenerator(
+                UUIDGenerator(),
+                config,
+                snowflakeColumnManager,
+            )
+        val snowflakeFinalTableNameGenerator =
+            SnowflakeTableSchemaMapper(
+                config = config,
+                tempTableNameGenerator = DefaultTempTableNameGenerator(),
+            )
         val dataSource =
             SnowflakeBeanFactory()
                 .snowflakeDataSource(snowflakeConfiguration = config, airbyteEdition = "COMMUNITY")
@@ -37,11 +50,11 @@ class SnowflakeRawDataDumper(
             ds.connection.use { connection ->
                 val statement = connection.createStatement()
                 val tableName =
-                    snowflakeFinalTableNameGenerator.getTableName(stream.mappedDescriptor)
+                    snowflakeFinalTableNameGenerator.toFinalTableName(stream.mappedDescriptor)
 
                 val resultSet =
                     statement.executeQuery(
-                        "SELECT * FROM ${sqlUtils.fullyQualifiedName(tableName)}"
+                        "SELECT * FROM ${sqlGenerator.fullyQualifiedName(tableName)}"
                     )
 
                 while (resultSet.next()) {

airbyte-integrations/connectors/destination-snowflake/src/test/kotlin/io/airbyte/integrations/destination/snowflake/SnowflakeBeanFactoryTest.kt

@@ -6,7 +6,7 @@ package io.airbyte.integrations.destination.snowflake
 
 import com.zaxxer.hikari.HikariConfig
 import com.zaxxer.hikari.HikariDataSource
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
+import io.airbyte.integrations.destination.snowflake.schema.toSnowflakeCompatibleName
 import io.airbyte.integrations.destination.snowflake.spec.CdcDeletionMode
 import io.airbyte.integrations.destination.snowflake.spec.KeyPairAuthConfiguration
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration

airbyte-integrations/connectors/destination-snowflake/src/test/kotlin/io/airbyte/integrations/destination/snowflake/check/SnowflakeCheckerTest.kt

@@ -5,11 +5,9 @@
 package io.airbyte.integrations.destination.snowflake.check
 
 import io.airbyte.integrations.destination.snowflake.client.SnowflakeAirbyteClient
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
+import io.airbyte.integrations.destination.snowflake.schema.SnowflakeColumnManager
+import io.airbyte.integrations.destination.snowflake.schema.toSnowflakeCompatibleName
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
-import io.airbyte.integrations.destination.snowflake.sql.DEFAULT_COLUMNS
-import io.airbyte.integrations.destination.snowflake.sql.RAW_DATA_COLUMN
-import io.airbyte.integrations.destination.snowflake.sql.SnowflakeColumnUtils
 import io.mockk.coEvery
 import io.mockk.coVerify
 import io.mockk.every
@@ -23,47 +21,31 @@ internal class SnowflakeCheckerTest {
     @ParameterizedTest
     @ValueSource(booleans = [true, false])
     fun testSuccessfulCheck(isLegacyRawTablesOnly: Boolean) {
-        val defaultColumnsMap =
-            if (isLegacyRawTablesOnly) {
-                linkedMapOf<String, String>().also { map ->
-                    (DEFAULT_COLUMNS + RAW_DATA_COLUMN).forEach {
-                        map[it.columnName] = it.columnType
-                    }
-                }
-            } else {
-                linkedMapOf<String, String>().also { map ->
-                    (DEFAULT_COLUMNS + RAW_DATA_COLUMN).forEach {
-                        map[it.columnName.toSnowflakeCompatibleName()] = it.columnType
-                    }
-                }
-            }
-        val defaultColumns = defaultColumnsMap.keys.toMutableList()
         val snowflakeAirbyteClient: SnowflakeAirbyteClient =
-            mockk(relaxed = true) {
-                coEvery { countTable(any()) } returns 1L
-                coEvery { describeTable(any()) } returns defaultColumnsMap
-            }
+            mockk(relaxed = true) { coEvery { countTable(any()) } returns 1L }
 
         val testSchema = "test-schema"
         val snowflakeConfiguration: SnowflakeConfiguration = mockk {
             every { schema } returns testSchema
             every { legacyRawTablesOnly } returns isLegacyRawTablesOnly
         }
-        val snowflakeColumnUtils =
-            mockk<SnowflakeColumnUtils>(relaxUnitFun = true) {
-                every { getFormattedDefaultColumnNames(any()) } returns defaultColumns
-            }
+
+        val columnManager: SnowflakeColumnManager = SnowflakeColumnManager(snowflakeConfiguration)
 
         val checker =
             SnowflakeChecker(
                 snowflakeAirbyteClient = snowflakeAirbyteClient,
                 snowflakeConfiguration = snowflakeConfiguration,
-                snowflakeColumnUtils = snowflakeColumnUtils,
+                columnManager = columnManager,
             )
         checker.check()
 
         coVerify(exactly = 1) {
-            snowflakeAirbyteClient.createNamespace(testSchema.toSnowflakeCompatibleName())
+            if (isLegacyRawTablesOnly) {
+                snowflakeAirbyteClient.createNamespace(testSchema)
+            } else {
+                snowflakeAirbyteClient.createNamespace(testSchema.toSnowflakeCompatibleName())
+            }
         }
         coVerify(exactly = 1) { snowflakeAirbyteClient.createTable(any(), any(), any(), any()) }
         coVerify(exactly = 1) { snowflakeAirbyteClient.dropTable(any()) }
@@ -72,48 +54,32 @@ internal class SnowflakeCheckerTest {
     @ParameterizedTest
     @ValueSource(booleans = [true, false])
     fun testUnsuccessfulCheck(isLegacyRawTablesOnly: Boolean) {
-        val defaultColumnsMap =
-            if (isLegacyRawTablesOnly) {
-                linkedMapOf<String, String>().also { map ->
-                    (DEFAULT_COLUMNS + RAW_DATA_COLUMN).forEach {
-                        map[it.columnName] = it.columnType
-                    }
-                }
-            } else {
-                linkedMapOf<String, String>().also { map ->
-                    (DEFAULT_COLUMNS + RAW_DATA_COLUMN).forEach {
-                        map[it.columnName.toSnowflakeCompatibleName()] = it.columnType
-                    }
-                }
-            }
-        val defaultColumns = defaultColumnsMap.keys.toMutableList()
         val snowflakeAirbyteClient: SnowflakeAirbyteClient =
-            mockk(relaxed = true) {
-                coEvery { countTable(any()) } returns 0L
-                coEvery { describeTable(any()) } returns defaultColumnsMap
-            }
+            mockk(relaxed = true) { coEvery { countTable(any()) } returns 0L }
 
         val testSchema = "test-schema"
         val snowflakeConfiguration: SnowflakeConfiguration = mockk {
             every { schema } returns testSchema
             every { legacyRawTablesOnly } returns isLegacyRawTablesOnly
         }
-        val snowflakeColumnUtils =
-            mockk<SnowflakeColumnUtils>(relaxUnitFun = true) {
-                every { getFormattedDefaultColumnNames(any()) } returns defaultColumns
-            }
+
+        val columnManager: SnowflakeColumnManager = SnowflakeColumnManager(snowflakeConfiguration)
 
         val checker =
             SnowflakeChecker(
                 snowflakeAirbyteClient = snowflakeAirbyteClient,
                 snowflakeConfiguration = snowflakeConfiguration,
-                snowflakeColumnUtils = snowflakeColumnUtils,
+                columnManager = columnManager,
             )
 
         assertThrows<IllegalArgumentException> { checker.check() }
 
         coVerify(exactly = 1) {
-            snowflakeAirbyteClient.createNamespace(testSchema.toSnowflakeCompatibleName())
+            if (isLegacyRawTablesOnly) {
+                snowflakeAirbyteClient.createNamespace(testSchema)
+            } else {
+                snowflakeAirbyteClient.createNamespace(testSchema.toSnowflakeCompatibleName())
+            }
         }
         coVerify(exactly = 1) { snowflakeAirbyteClient.createTable(any(), any(), any(), any()) }
         coVerify(exactly = 1) { snowflakeAirbyteClient.dropTable(any()) }

airbyte-integrations/connectors/destination-snowflake/src/test/kotlin/io/airbyte/integrations/destination/snowflake/client/SnowflakeAirbyteClientTest.kt

@@ -6,24 +6,16 @@ package io.airbyte.integrations.destination.snowflake.client
 
 import io.airbyte.cdk.ConfigErrorException
 import io.airbyte.cdk.load.command.DestinationStream
-import io.airbyte.cdk.load.command.NamespaceMapper
-import io.airbyte.cdk.load.command.Overwrite
 import io.airbyte.cdk.load.component.ColumnType
-import io.airbyte.cdk.load.config.NamespaceDefinitionType
-import io.airbyte.cdk.load.data.AirbyteType
-import io.airbyte.cdk.load.data.FieldType
 import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_GENERATION_ID
 import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_RAW_ID
+import io.airbyte.cdk.load.schema.model.TableName
 import io.airbyte.cdk.load.table.ColumnNameMapping
-import io.airbyte.cdk.load.table.TableName
-import io.airbyte.integrations.destination.snowflake.db.ColumnDefinition
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
+import io.airbyte.integrations.destination.snowflake.schema.SnowflakeColumnManager
+import io.airbyte.integrations.destination.snowflake.schema.toSnowflakeCompatibleName
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
 import io.airbyte.integrations.destination.snowflake.sql.COUNT_TOTAL_ALIAS
-import io.airbyte.integrations.destination.snowflake.sql.ColumnAndType
-import io.airbyte.integrations.destination.snowflake.sql.DEFAULT_COLUMNS
 import io.airbyte.integrations.destination.snowflake.sql.QUOTE
-import io.airbyte.integrations.destination.snowflake.sql.SnowflakeColumnUtils
 import io.airbyte.integrations.destination.snowflake.sql.SnowflakeDirectLoadSqlGenerator
 import io.mockk.Runs
 import io.mockk.every
@@ -49,31 +41,18 @@ internal class SnowflakeAirbyteClientTest {
     private lateinit var client: SnowflakeAirbyteClient
     private lateinit var dataSource: DataSource
     private lateinit var sqlGenerator: SnowflakeDirectLoadSqlGenerator
-    private lateinit var snowflakeColumnUtils: SnowflakeColumnUtils
     private lateinit var snowflakeConfiguration: SnowflakeConfiguration
+    private lateinit var columnManager: SnowflakeColumnManager
 
     @BeforeEach
     fun setup() {
         dataSource = mockk()
         sqlGenerator = mockk(relaxed = true)
-        snowflakeColumnUtils =
-            mockk(relaxed = true) {
-                every { formatColumnName(any()) } answers
-                    {
-                        firstArg<String>().toSnowflakeCompatibleName()
-                    }
-                every { getFormattedDefaultColumnNames(any()) } returns
-                    DEFAULT_COLUMNS.map { it.columnName.toSnowflakeCompatibleName() }
-            }
         snowflakeConfiguration =
             mockk(relaxed = true) { every { database } returns "test_database" }
+        columnManager = mockk(relaxed = true)
         client =
-            SnowflakeAirbyteClient(
-                dataSource,
-                sqlGenerator,
-                snowflakeColumnUtils,
-                snowflakeConfiguration
-            )
+            SnowflakeAirbyteClient(dataSource, sqlGenerator, snowflakeConfiguration, columnManager)
     }
 
     @Test
@@ -231,7 +210,7 @@ internal class SnowflakeAirbyteClientTest {
     @Test
     fun testCreateTable() {
         val columnNameMapping = mockk<ColumnNameMapping>(relaxed = true)
-        val stream = mockk<DestinationStream>()
+        val stream = mockk<DestinationStream>(relaxed = true)
         val tableName = TableName(namespace = "namespace", name = "name")
         val resultSet = mockk<ResultSet>(relaxed = true)
         val statement =
@@ -254,9 +233,7 @@ internal class SnowflakeAirbyteClientTest {
                 columnNameMapping = columnNameMapping,
                 replace = true,
             )
-            verify(exactly = 1) {
-                sqlGenerator.createTable(stream, tableName, columnNameMapping, true)
-            }
+            verify(exactly = 1) { sqlGenerator.createTable(tableName, any(), true) }
             verify(exactly = 1) { sqlGenerator.createSnowflakeStage(tableName) }
             verify(exactly = 2) { mockConnection.close() }
         }
@@ -288,7 +265,7 @@ internal class SnowflakeAirbyteClientTest {
                 targetTableName = destinationTableName,
             )
             verify(exactly = 1) {
-                sqlGenerator.copyTable(columnNameMapping, sourceTableName, destinationTableName)
+                sqlGenerator.copyTable(any<Set<String>>(), sourceTableName, destinationTableName)
             }
             verify(exactly = 1) { mockConnection.close() }
         }
@@ -299,7 +276,7 @@ internal class SnowflakeAirbyteClientTest {
         val columnNameMapping = mockk<ColumnNameMapping>(relaxed = true)
         val sourceTableName = TableName(namespace = "namespace", name = "source")
         val destinationTableName = TableName(namespace = "namespace", name = "destination")
-        val stream = mockk<DestinationStream>()
+        val stream = mockk<DestinationStream>(relaxed = true)
         val resultSet = mockk<ResultSet>(relaxed = true)
         val statement =
             mockk<Statement> {
@@ -322,12 +299,7 @@ internal class SnowflakeAirbyteClientTest {
                 targetTableName = destinationTableName,
             )
             verify(exactly = 1) {
-                sqlGenerator.upsertTable(
-                    stream,
-                    columnNameMapping,
-                    sourceTableName,
-                    destinationTableName
-                )
+                sqlGenerator.upsertTable(any(), sourceTableName, destinationTableName)
             }
             verify(exactly = 1) { mockConnection.close() }
         }
@@ -379,7 +351,7 @@ internal class SnowflakeAirbyteClientTest {
             }
 
         every { dataSource.connection } returns mockConnection
-        every { snowflakeColumnUtils.getGenerationIdColumnName() } returns generationIdColumnName
+        every { columnManager.getGenerationIdColumnName() } returns generationIdColumnName
         every { sqlGenerator.getGenerationId(tableName) } returns
             "SELECT $generationIdColumnName FROM ${tableName.toPrettyString(QUOTE)}"
 
@@ -501,8 +473,8 @@ internal class SnowflakeAirbyteClientTest {
         every { dataSource.connection } returns mockConnection
 
         runBlocking {
-            client.copyFromStage(tableName, "test.csv.gz")
-            verify(exactly = 1) { sqlGenerator.copyFromStage(tableName, "test.csv.gz") }
+            client.copyFromStage(tableName, "test.csv.gz", listOf())
+            verify(exactly = 1) { sqlGenerator.copyFromStage(tableName, "test.csv.gz", listOf()) }
             verify(exactly = 1) { mockConnection.close() }
         }
     }
@@ -556,7 +528,7 @@ internal class SnowflakeAirbyteClientTest {
             "COL1" andThen
             COLUMN_NAME_AB_RAW_ID.toSnowflakeCompatibleName() andThen
             "COL2"
-        every { resultSet.getString("type") } returns "VARCHAR(255)" andThen "NUMBER(38,0)"
+        every { resultSet.getString("type") } returns "VARCHAR(255)" andThen "NUMBER"
         every { resultSet.getString("null?") } returns "Y" andThen "N" andThen "N"
 
         val statement =
@@ -571,6 +543,10 @@ internal class SnowflakeAirbyteClientTest {
 
         every { dataSource.connection } returns connection
 
+        // Mock the columnManager to return the correct set of meta columns
+        every { columnManager.getMetaColumnNames() } returns
+            setOf(COLUMN_NAME_AB_RAW_ID.toSnowflakeCompatibleName())
+
         val result = client.getColumnsFromDb(tableName)
 
         val expectedColumns =
@@ -582,81 +558,6 @@ internal class SnowflakeAirbyteClientTest {
         assertEquals(expectedColumns, result)
     }
 
-    @Test
-    fun `getColumnsFromStream should return correct column definitions`() {
-        val schema = mockk<AirbyteType>()
-        val stream =
-            DestinationStream(
-                unmappedNamespace = "test_namespace",
-                unmappedName = "test_stream",
-                importType = Overwrite,
-                schema = schema,
-                generationId = 1,
-                minimumGenerationId = 1,
-                syncId = 1,
-                namespaceMapper = NamespaceMapper(NamespaceDefinitionType.DESTINATION)
-            )
-        val columnNameMapping =
-            ColumnNameMapping(
-                mapOf(
-                    "col1" to "COL1_MAPPED",
-                    "col2" to "COL2_MAPPED",
-                )
-            )
-
-        val col1FieldType = mockk<FieldType>()
-        every { col1FieldType.type } returns mockk()
-
-        val col2FieldType = mockk<FieldType>()
-        every { col2FieldType.type } returns mockk()
-
-        every { schema.asColumns() } returns
-            linkedMapOf("col1" to col1FieldType, "col2" to col2FieldType)
-        every { snowflakeColumnUtils.toDialectType(col1FieldType.type) } returns "VARCHAR(255)"
-        every { snowflakeColumnUtils.toDialectType(col2FieldType.type) } returns "NUMBER(38,0)"
-        every { snowflakeColumnUtils.columnsAndTypes(any(), any()) } returns
-            listOf(ColumnAndType("COL1_MAPPED", "VARCHAR"), ColumnAndType("COL2_MAPPED", "NUMBER"))
-        every { snowflakeColumnUtils.formatColumnName(any(), false) } answers
-            {
-                firstArg<String>().toSnowflakeCompatibleName()
-            }
-
-        val result = client.getColumnsFromStream(stream, columnNameMapping)
-
-        val expectedColumns =
-            mapOf(
-                "COL1_MAPPED" to ColumnType("VARCHAR", true),
-                "COL2_MAPPED" to ColumnType("NUMBER", true),
-            )
-
-        assertEquals(expectedColumns, result)
-    }
-
-    @Test
-    fun `generateSchemaChanges should correctly identify changes`() {
-        val columnsInDb =
-            setOf(
-                ColumnDefinition("COL1", "VARCHAR"),
-                ColumnDefinition("COL2", "NUMBER"),
-                ColumnDefinition("COL3", "BOOLEAN")
-            )
-        val columnsInStream =
-            setOf(
-                ColumnDefinition("COL1", "VARCHAR"), // Unchanged
-                ColumnDefinition("COL3", "TEXT"), // Modified
-                ColumnDefinition("COL4", "DATE") // Added
-            )
-
-        val (added, deleted, modified) = client.generateSchemaChanges(columnsInDb, columnsInStream)
-
-        assertEquals(1, added.size)
-        assertEquals("COL4", added.first().name)
-        assertEquals(1, deleted.size)
-        assertEquals("COL2", deleted.first().name)
-        assertEquals(1, modified.size)
-        assertEquals("COL3", modified.first().name)
-    }
-
     @Test
     fun testCreateNamespaceWithNetworkFailure() {
         val namespace = "test_namespace"

airbyte-integrations/connectors/destination-snowflake/src/test/kotlin/io/airbyte/integrations/destination/snowflake/dataflow/SnowflakeAggregateFactoryTest.kt

@@ -4,14 +4,19 @@
 
 package io.airbyte.integrations.destination.snowflake.dataflow
 
+import io.airbyte.cdk.load.command.DestinationCatalog
+import io.airbyte.cdk.load.command.DestinationStream
 import io.airbyte.cdk.load.command.DestinationStream.Descriptor
 import io.airbyte.cdk.load.dataflow.aggregate.StoreKey
-import io.airbyte.cdk.load.orchestration.db.direct_load_table.DirectLoadTableExecutionConfig
-import io.airbyte.cdk.load.table.TableName
+import io.airbyte.cdk.load.schema.model.TableName
+import io.airbyte.cdk.load.table.directload.DirectLoadTableExecutionConfig
 import io.airbyte.cdk.load.write.StreamStateStore
 import io.airbyte.integrations.destination.snowflake.client.SnowflakeAirbyteClient
+import io.airbyte.integrations.destination.snowflake.schema.SnowflakeColumnManager
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
-import io.airbyte.integrations.destination.snowflake.sql.SnowflakeColumnUtils
+import io.airbyte.integrations.destination.snowflake.write.load.SnowflakeRawRecordFormatter
+import io.airbyte.integrations.destination.snowflake.write.load.SnowflakeRecordFormatter
+import io.airbyte.integrations.destination.snowflake.write.load.SnowflakeSchemaRecordFormatter
 import io.mockk.every
 import io.mockk.mockk
 import org.junit.jupiter.api.Assertions.assertEquals
@@ -22,27 +27,33 @@ internal class SnowflakeAggregateFactoryTest {
     @Test
     fun testCreatingAggregateWithRawBuffer() {
         val descriptor = Descriptor(namespace = "namespace", name = "name")
-        val directLoadTableExecutionConfig =
-            DirectLoadTableExecutionConfig(
-                tableName =
-                    TableName(
-                        namespace = descriptor.namespace!!,
-                        name = descriptor.name,
-                    )
+        val tableName =
+            TableName(
+                namespace = descriptor.namespace!!,
+                name = descriptor.name,
             )
+        val directLoadTableExecutionConfig = DirectLoadTableExecutionConfig(tableName = tableName)
         val key = StoreKey(namespace = descriptor.namespace!!, name = descriptor.name)
         val streamStore = StreamStateStore<DirectLoadTableExecutionConfig>()
-        streamStore.put(descriptor, directLoadTableExecutionConfig)
+        streamStore.put(key, directLoadTableExecutionConfig)
+
+        val stream = mockk<DestinationStream>(relaxed = true)
+        val catalog = mockk<DestinationCatalog> { every { getStream(key) } returns stream }
+
         val snowflakeClient = mockk<SnowflakeAirbyteClient>(relaxed = true)
         val snowflakeConfiguration =
             mockk<SnowflakeConfiguration> { every { legacyRawTablesOnly } returns true }
-        val snowflakeColumnUtils = mockk<SnowflakeColumnUtils>(relaxed = true)
+        val columnManager = SnowflakeColumnManager(snowflakeConfiguration)
+        val snowflakeRecordFormatter: SnowflakeRecordFormatter = SnowflakeRawRecordFormatter()
+
         val factory =
             SnowflakeAggregateFactory(
                 snowflakeClient = snowflakeClient,
                 streamStateStore = streamStore,
                 snowflakeConfiguration = snowflakeConfiguration,
-                snowflakeColumnUtils = snowflakeColumnUtils,
+                catalog = catalog,
+                columnManager = columnManager,
+                snowflakeRecordFormatter = snowflakeRecordFormatter,
             )
         val aggregate = factory.create(key)
         assertNotNull(aggregate)
@@ -52,26 +63,33 @@ internal class SnowflakeAggregateFactoryTest {
     @Test
     fun testCreatingAggregateWithStagingBuffer() {
         val descriptor = Descriptor(namespace = "namespace", name = "name")
-        val directLoadTableExecutionConfig =
-            DirectLoadTableExecutionConfig(
-                tableName =
-                    TableName(
-                        namespace = descriptor.namespace!!,
-                        name = descriptor.name,
-                    )
+        val tableName =
+            TableName(
+                namespace = descriptor.namespace!!,
+                name = descriptor.name,
             )
+        val directLoadTableExecutionConfig = DirectLoadTableExecutionConfig(tableName = tableName)
         val key = StoreKey(namespace = descriptor.namespace!!, name = descriptor.name)
         val streamStore = StreamStateStore<DirectLoadTableExecutionConfig>()
-        streamStore.put(descriptor, directLoadTableExecutionConfig)
+        streamStore.put(key, directLoadTableExecutionConfig)
+
+        val stream = mockk<DestinationStream>(relaxed = true)
+        val catalog = mockk<DestinationCatalog> { every { getStream(key) } returns stream }
+
         val snowflakeClient = mockk<SnowflakeAirbyteClient>(relaxed = true)
-        val snowflakeConfiguration = mockk<SnowflakeConfiguration>(relaxed = true)
-        val snowflakeColumnUtils = mockk<SnowflakeColumnUtils>(relaxed = true)
+        val snowflakeConfiguration =
+            mockk<SnowflakeConfiguration> { every { legacyRawTablesOnly } returns false }
+        val columnManager = SnowflakeColumnManager(snowflakeConfiguration)
+        val snowflakeRecordFormatter: SnowflakeRecordFormatter = SnowflakeSchemaRecordFormatter()
+
         val factory =
             SnowflakeAggregateFactory(
                 snowflakeClient = snowflakeClient,
                 streamStateStore = streamStore,
                 snowflakeConfiguration = snowflakeConfiguration,
-                snowflakeColumnUtils = snowflakeColumnUtils,
+                catalog = catalog,
+                columnManager = columnManager,
+                snowflakeRecordFormatter = snowflakeRecordFormatter,
             )
         val aggregate = factory.create(key)
         assertNotNull(aggregate)

airbyte-integrations/connectors/destination-snowflake/src/test/kotlin/io/airbyte/integrations/destination/snowflake/db/SnowflakeColumnNameGeneratorTest.kt

@@ -1,23 +0,0 @@
-/*
- * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
- */
-
-package io.airbyte.integrations.destination.snowflake.db
-
-import io.mockk.every
-import io.mockk.mockk
-import org.junit.jupiter.api.Assertions.assertEquals
-import org.junit.jupiter.api.Test
-
-internal class SnowflakeColumnNameGeneratorTest {
-
-    @Test
-    fun testGetColumnName() {
-        val column = "test-column"
-        val generator =
-            SnowflakeColumnNameGenerator(mockk { every { legacyRawTablesOnly } returns false })
-        val columnName = generator.getColumnName(column)
-        assertEquals(column.toSnowflakeCompatibleName(), columnName.displayName)
-        assertEquals(column.toSnowflakeCompatibleName(), columnName.canonicalName)
-    }
-}

airbyte-integrations/connectors/destination-snowflake/src/test/kotlin/io/airbyte/integrations/destination/snowflake/db/SnowflakeFinalTableNameGeneratorTest.kt

@@ -1,72 +0,0 @@
-/*
- * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
- */
-
-package io.airbyte.integrations.destination.snowflake.db
-
-import io.airbyte.cdk.load.command.DestinationStream
-import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
-import io.mockk.every
-import io.mockk.mockk
-import org.junit.jupiter.api.Assertions.assertEquals
-import org.junit.jupiter.api.Test
-
-internal class SnowflakeFinalTableNameGeneratorTest {
-
-    @Test
-    fun testGetTableNameWithInternalNamespace() {
-        val configuration =
-            mockk<SnowflakeConfiguration> {
-                every { internalTableSchema } returns "test-internal-namespace"
-                every { legacyRawTablesOnly } returns true
-            }
-        val generator = SnowflakeFinalTableNameGenerator(config = configuration)
-        val streamName = "test-stream-name"
-        val streamNamespace = "test-stream-namespace"
-        val streamDescriptor =
-            mockk<DestinationStream.Descriptor> {
-                every { namespace } returns streamNamespace
-                every { name } returns streamName
-            }
-        val tableName = generator.getTableName(streamDescriptor)
-        assertEquals("test-stream-namespace_raw__stream_test-stream-name", tableName.name)
-        assertEquals("test-internal-namespace", tableName.namespace)
-    }
-
-    @Test
-    fun testGetTableNameWithNamespace() {
-        val configuration =
-            mockk<SnowflakeConfiguration> { every { legacyRawTablesOnly } returns false }
-        val generator = SnowflakeFinalTableNameGenerator(config = configuration)
-        val streamName = "test-stream-name"
-        val streamNamespace = "test-stream-namespace"
-        val streamDescriptor =
-            mockk<DestinationStream.Descriptor> {
-                every { namespace } returns streamNamespace
-                every { name } returns streamName
-            }
-        val tableName = generator.getTableName(streamDescriptor)
-        assertEquals("TEST-STREAM-NAME", tableName.name)
-        assertEquals("TEST-STREAM-NAMESPACE", tableName.namespace)
-    }
-
-    @Test
-    fun testGetTableNameWithDefaultNamespace() {
-        val defaultNamespace = "test-default-namespace"
-        val configuration =
-            mockk<SnowflakeConfiguration> {
-                every { schema } returns defaultNamespace
-                every { legacyRawTablesOnly } returns false
-            }
-        val generator = SnowflakeFinalTableNameGenerator(config = configuration)
-        val streamName = "test-stream-name"
-        val streamDescriptor =
-            mockk<DestinationStream.Descriptor> {
-                every { namespace } returns null
-                every { name } returns streamName
-            }
-        val tableName = generator.getTableName(streamDescriptor)
-        assertEquals("TEST-STREAM-NAME", tableName.name)
-        assertEquals("TEST-DEFAULT-NAMESPACE", tableName.namespace)
-    }
-}

airbyte-integrations/connectors/destination-snowflake/src/test/kotlin/io/airbyte/integrations/destination/snowflake/db/SnowflakeNameGeneratorsTest.kt

@@ -1,27 +0,0 @@
-/*
- * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
- */
-
-package io.airbyte.integrations.destination.snowflake.db
-
-import org.junit.jupiter.api.Assertions.assertEquals
-import org.junit.jupiter.params.ParameterizedTest
-import org.junit.jupiter.params.provider.CsvSource
-
-internal class SnowflakeNameGeneratorsTest {
-
-    @ParameterizedTest
-    @CsvSource(
-        value =
-            [
-                "test-name,TEST-NAME",
-                "1-test-name,1-TEST-NAME",
-                "test-name!!!,TEST-NAME!!!",
-                "test\${name,TEST__NAME",
-                "test\"name,TEST\"\"NAME",
-            ]
-    )
-    fun testToSnowflakeCompatibleName(name: String, expected: String) {
-        assertEquals(expected, name.toSnowflakeCompatibleName())
-    }
-}

airbyte-integrations/connectors/destination-snowflake/src/test/kotlin/io/airbyte/integrations/destination/snowflake/sql/SnowflakeColumnUtilsTest.kt

@@ -1,358 +0,0 @@
-/*
- * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
- */
-
-package io.airbyte.integrations.destination.snowflake.sql
-
-import com.fasterxml.jackson.databind.JsonNode
-import io.airbyte.cdk.load.data.ArrayType
-import io.airbyte.cdk.load.data.ArrayTypeWithoutSchema
-import io.airbyte.cdk.load.data.BooleanType
-import io.airbyte.cdk.load.data.DateType
-import io.airbyte.cdk.load.data.FieldType
-import io.airbyte.cdk.load.data.IntegerType
-import io.airbyte.cdk.load.data.NumberType
-import io.airbyte.cdk.load.data.ObjectType
-import io.airbyte.cdk.load.data.ObjectTypeWithEmptySchema
-import io.airbyte.cdk.load.data.ObjectTypeWithoutSchema
-import io.airbyte.cdk.load.data.StringType
-import io.airbyte.cdk.load.data.TimeTypeWithTimezone
-import io.airbyte.cdk.load.data.TimeTypeWithoutTimezone
-import io.airbyte.cdk.load.data.TimestampTypeWithTimezone
-import io.airbyte.cdk.load.data.TimestampTypeWithoutTimezone
-import io.airbyte.cdk.load.data.UnionType
-import io.airbyte.cdk.load.data.UnknownType
-import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_DATA
-import io.airbyte.cdk.load.orchestration.db.ColumnNameGenerator
-import io.airbyte.cdk.load.table.CDC_DELETED_AT_COLUMN
-import io.airbyte.cdk.load.table.ColumnNameMapping
-import io.airbyte.integrations.destination.snowflake.db.SnowflakeColumnNameGenerator
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
-import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
-import io.mockk.every
-import io.mockk.mockk
-import kotlin.collections.LinkedHashMap
-import org.junit.jupiter.api.Assertions.assertEquals
-import org.junit.jupiter.api.BeforeEach
-import org.junit.jupiter.api.Test
-import org.junit.jupiter.params.ParameterizedTest
-import org.junit.jupiter.params.provider.CsvSource
-
-internal class SnowflakeColumnUtilsTest {
-
-    private lateinit var snowflakeConfiguration: SnowflakeConfiguration
-    private lateinit var snowflakeColumnUtils: SnowflakeColumnUtils
-    private lateinit var snowflakeColumnNameGenerator: SnowflakeColumnNameGenerator
-
-    @BeforeEach
-    fun setup() {
-        snowflakeConfiguration = mockk(relaxed = true)
-        snowflakeColumnNameGenerator =
-            mockk(relaxed = true) {
-                every { getColumnName(any()) } answers
-                    {
-                        val displayName =
-                            if (snowflakeConfiguration.legacyRawTablesOnly) firstArg<String>()
-                            else firstArg<String>().toSnowflakeCompatibleName()
-                        val canonicalName =
-                            if (snowflakeConfiguration.legacyRawTablesOnly) firstArg<String>()
-                            else firstArg<String>().toSnowflakeCompatibleName()
-                        ColumnNameGenerator.ColumnName(
-                            displayName = displayName,
-                            canonicalName = canonicalName,
-                        )
-                    }
-            }
-        snowflakeColumnUtils =
-            SnowflakeColumnUtils(snowflakeConfiguration, snowflakeColumnNameGenerator)
-    }
-
-    @Test
-    fun testDefaultColumns() {
-        val expectedDefaultColumns = DEFAULT_COLUMNS
-        assertEquals(expectedDefaultColumns, snowflakeColumnUtils.defaultColumns())
-    }
-
-    @Test
-    fun testDefaultRawColumns() {
-        every { snowflakeConfiguration.legacyRawTablesOnly } returns true
-
-        val expectedDefaultColumns = DEFAULT_COLUMNS + RAW_COLUMNS
-
-        assertEquals(expectedDefaultColumns, snowflakeColumnUtils.defaultColumns())
-    }
-
-    @Test
-    fun testGetFormattedDefaultColumnNames() {
-        val expectedDefaultColumnNames =
-            DEFAULT_COLUMNS.map { it.columnName.toSnowflakeCompatibleName() }
-        val defaultColumnNames = snowflakeColumnUtils.getFormattedDefaultColumnNames()
-        assertEquals(expectedDefaultColumnNames, defaultColumnNames)
-    }
-
-    @Test
-    fun testGetFormattedDefaultColumnNamesQuoted() {
-        val expectedDefaultColumnNames =
-            DEFAULT_COLUMNS.map { it.columnName.toSnowflakeCompatibleName().quote() }
-        val defaultColumnNames = snowflakeColumnUtils.getFormattedDefaultColumnNames(true)
-        assertEquals(expectedDefaultColumnNames, defaultColumnNames)
-    }
-
-    @Test
-    fun testGetColumnName() {
-        val columnNameMapping = ColumnNameMapping(mapOf("original" to "actual"))
-        val columnNames = snowflakeColumnUtils.getColumnNames(columnNameMapping)
-        val expectedColumnNames =
-            (DEFAULT_COLUMNS.map { it.columnName.toSnowflakeCompatibleName() } + listOf("actual"))
-                .joinToString(",") { it.quote() }
-        assertEquals(expectedColumnNames, columnNames)
-    }
-
-    @Test
-    fun testGetRawColumnName() {
-        every { snowflakeConfiguration.legacyRawTablesOnly } returns true
-
-        val columnNameMapping = ColumnNameMapping(mapOf("original" to "actual"))
-        val columnNames = snowflakeColumnUtils.getColumnNames(columnNameMapping)
-        val expectedColumnNames =
-            (DEFAULT_COLUMNS.map { it.columnName } + RAW_COLUMNS.map { it.columnName })
-                .joinToString(",") { it.quote() }
-        assertEquals(expectedColumnNames, columnNames)
-    }
-
-    @Test
-    fun testGetRawFormattedColumnNames() {
-        every { snowflakeConfiguration.legacyRawTablesOnly } returns true
-        val columnNameMapping = ColumnNameMapping(mapOf("original" to "actual"))
-        val schemaColumns =
-            mapOf(
-                "column_one" to FieldType(StringType, true),
-                "column_two" to FieldType(IntegerType, true),
-                "original" to FieldType(StringType, true),
-                CDC_DELETED_AT_COLUMN to FieldType(TimestampTypeWithTimezone, true)
-            )
-        val expectedColumnNames =
-            DEFAULT_COLUMNS.map { it.columnName.quote() } +
-                RAW_COLUMNS.map { it.columnName.quote() }
-
-        val columnNames =
-            snowflakeColumnUtils.getFormattedColumnNames(
-                columns = schemaColumns,
-                columnNameMapping = columnNameMapping
-            )
-        assertEquals(expectedColumnNames.size, columnNames.size)
-        assertEquals(expectedColumnNames.sorted(), columnNames.sorted())
-    }
-
-    @Test
-    fun testGetFormattedColumnNames() {
-        val columnNameMapping = ColumnNameMapping(mapOf("original" to "actual"))
-        val schemaColumns =
-            mapOf(
-                "column_one" to FieldType(StringType, true),
-                "column_two" to FieldType(IntegerType, true),
-                "original" to FieldType(StringType, true),
-                CDC_DELETED_AT_COLUMN to FieldType(TimestampTypeWithTimezone, true)
-            )
-        val expectedColumnNames =
-            listOf(
-                    "actual",
-                    "column_one",
-                    "column_two",
-                    CDC_DELETED_AT_COLUMN,
-                )
-                .map { it.quote() } +
-                DEFAULT_COLUMNS.map { it.columnName.toSnowflakeCompatibleName().quote() }
-        val columnNames =
-            snowflakeColumnUtils.getFormattedColumnNames(
-                columns = schemaColumns,
-                columnNameMapping = columnNameMapping
-            )
-        assertEquals(expectedColumnNames.size, columnNames.size)
-        assertEquals(expectedColumnNames.sorted(), columnNames.sorted())
-    }
-
-    @Test
-    fun testGetFormattedColumnNamesNoQuotes() {
-        val columnNameMapping = ColumnNameMapping(mapOf("original" to "actual"))
-        val schemaColumns =
-            mapOf(
-                "column_one" to FieldType(StringType, true),
-                "column_two" to FieldType(IntegerType, true),
-                "original" to FieldType(StringType, true),
-                CDC_DELETED_AT_COLUMN to FieldType(TimestampTypeWithTimezone, true)
-            )
-        val expectedColumnNames =
-            listOf(
-                "actual",
-                "column_one",
-                "column_two",
-                CDC_DELETED_AT_COLUMN,
-            ) + DEFAULT_COLUMNS.map { it.columnName.toSnowflakeCompatibleName() }
-        val columnNames =
-            snowflakeColumnUtils.getFormattedColumnNames(
-                columns = schemaColumns,
-                columnNameMapping = columnNameMapping,
-                quote = false
-            )
-        assertEquals(expectedColumnNames.size, columnNames.size)
-        assertEquals(expectedColumnNames.sorted(), columnNames.sorted())
-    }
-
-    @Test
-    fun testGeneratingRawTableColumnsAndTypesNoColumnMapping() {
-        every { snowflakeConfiguration.legacyRawTablesOnly } returns true
-
-        val columns =
-            snowflakeColumnUtils.columnsAndTypes(
-                columns = emptyMap(),
-                columnNameMapping = ColumnNameMapping(emptyMap())
-            )
-        assertEquals(DEFAULT_COLUMNS.size + RAW_COLUMNS.size, columns.size)
-        assertEquals(
-            "${SnowflakeDataType.VARIANT.typeName} $NOT_NULL",
-            columns.find { it.columnName == RAW_DATA_COLUMN.columnName }?.columnType
-        )
-    }
-
-    @Test
-    fun testGeneratingColumnsAndTypesNoColumnMapping() {
-        val columnName = "test-column"
-        val fieldType = FieldType(StringType, false)
-        val declaredColumns = mapOf(columnName to fieldType)
-
-        val columns =
-            snowflakeColumnUtils.columnsAndTypes(
-                columns = declaredColumns,
-                columnNameMapping = ColumnNameMapping(emptyMap())
-            )
-        assertEquals(DEFAULT_COLUMNS.size + 1, columns.size)
-        assertEquals(
-            "${SnowflakeDataType.VARCHAR.typeName} $NOT_NULL",
-            columns.find { it.columnName == columnName }?.columnType
-        )
-    }
-
-    @Test
-    fun testGeneratingColumnsAndTypesWithColumnMapping() {
-        val columnName = "test-column"
-        val mappedColumnName = "mapped-column-name"
-        val fieldType = FieldType(StringType, false)
-        val declaredColumns = mapOf(columnName to fieldType)
-        val columnNameMapping = ColumnNameMapping(mapOf(columnName to mappedColumnName))
-
-        val columns =
-            snowflakeColumnUtils.columnsAndTypes(
-                columns = declaredColumns,
-                columnNameMapping = columnNameMapping
-            )
-        assertEquals(DEFAULT_COLUMNS.size + 1, columns.size)
-        assertEquals(
-            "${SnowflakeDataType.VARCHAR.typeName} $NOT_NULL",
-            columns.find { it.columnName == mappedColumnName }?.columnType
-        )
-    }
-
-    @Test
-    fun testToDialectType() {
-        assertEquals(
-            SnowflakeDataType.BOOLEAN.typeName,
-            snowflakeColumnUtils.toDialectType(BooleanType)
-        )
-        assertEquals(SnowflakeDataType.DATE.typeName, snowflakeColumnUtils.toDialectType(DateType))
-        assertEquals(
-            SnowflakeDataType.NUMBER.typeName,
-            snowflakeColumnUtils.toDialectType(IntegerType)
-        )
-        assertEquals(
-            SnowflakeDataType.FLOAT.typeName,
-            snowflakeColumnUtils.toDialectType(NumberType)
-        )
-        assertEquals(
-            SnowflakeDataType.VARCHAR.typeName,
-            snowflakeColumnUtils.toDialectType(StringType)
-        )
-        assertEquals(
-            SnowflakeDataType.VARCHAR.typeName,
-            snowflakeColumnUtils.toDialectType(TimeTypeWithTimezone)
-        )
-        assertEquals(
-            SnowflakeDataType.TIME.typeName,
-            snowflakeColumnUtils.toDialectType(TimeTypeWithoutTimezone)
-        )
-        assertEquals(
-            SnowflakeDataType.TIMESTAMP_TZ.typeName,
-            snowflakeColumnUtils.toDialectType(TimestampTypeWithTimezone)
-        )
-        assertEquals(
-            SnowflakeDataType.TIMESTAMP_NTZ.typeName,
-            snowflakeColumnUtils.toDialectType(TimestampTypeWithoutTimezone)
-        )
-        assertEquals(
-            SnowflakeDataType.ARRAY.typeName,
-            snowflakeColumnUtils.toDialectType(ArrayType(items = FieldType(StringType, false)))
-        )
-        assertEquals(
-            SnowflakeDataType.ARRAY.typeName,
-            snowflakeColumnUtils.toDialectType(ArrayTypeWithoutSchema)
-        )
-        assertEquals(
-            SnowflakeDataType.OBJECT.typeName,
-            snowflakeColumnUtils.toDialectType(
-                ObjectType(
-                    properties = LinkedHashMap(),
-                    additionalProperties = false,
-                )
-            )
-        )
-        assertEquals(
-            SnowflakeDataType.OBJECT.typeName,
-            snowflakeColumnUtils.toDialectType(ObjectTypeWithEmptySchema)
-        )
-        assertEquals(
-            SnowflakeDataType.OBJECT.typeName,
-            snowflakeColumnUtils.toDialectType(ObjectTypeWithoutSchema)
-        )
-        assertEquals(
-            SnowflakeDataType.VARIANT.typeName,
-            snowflakeColumnUtils.toDialectType(
-                UnionType(
-                    options = setOf(StringType),
-                    isLegacyUnion = true,
-                )
-            )
-        )
-        assertEquals(
-            SnowflakeDataType.VARIANT.typeName,
-            snowflakeColumnUtils.toDialectType(
-                UnionType(
-                    options = emptySet(),
-                    isLegacyUnion = false,
-                )
-            )
-        )
-        assertEquals(
-            SnowflakeDataType.VARIANT.typeName,
-            snowflakeColumnUtils.toDialectType(UnknownType(schema = mockk<JsonNode>()))
-        )
-    }
-
-    @ParameterizedTest
-    @CsvSource(
-        value =
-            [
-                "$COLUMN_NAME_DATA, true, \"$COLUMN_NAME_DATA\"",
-                "some-other_Column, true, \"SOME-OTHER_COLUMN\"",
-                "$COLUMN_NAME_DATA, false, $COLUMN_NAME_DATA",
-                "some-other_Column, false, SOME-OTHER_COLUMN",
-                "$COLUMN_NAME_DATA, true, \"$COLUMN_NAME_DATA\"",
-                "some-other_Column, true, \"SOME-OTHER_COLUMN\"",
-            ]
-    )
-    fun testFormatColumnName(columnName: String, quote: Boolean, expectedFormattedName: String) {
-        assertEquals(
-            expectedFormattedName,
-            snowflakeColumnUtils.formatColumnName(columnName, quote)
-        )
-    }
-}

airbyte-integrations/connectors/destination-snowflake/src/test/kotlin/io/airbyte/integrations/destination/snowflake/sql/SnowflakeSqlNameUtilsTest.kt

@@ -1,97 +0,0 @@
-/*
- * Copyright (c) 2025 Airbyte, Inc., all rights reserved.
- */
-
-package io.airbyte.integrations.destination.snowflake.sql
-
-import io.airbyte.cdk.load.table.TableName
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
-import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
-import io.mockk.every
-import io.mockk.mockk
-import org.junit.jupiter.api.Assertions.assertEquals
-import org.junit.jupiter.api.BeforeEach
-import org.junit.jupiter.api.Test
-
-internal class SnowflakeSqlNameUtilsTest {
-
-    private lateinit var snowflakeConfiguration: SnowflakeConfiguration
-    private lateinit var snowflakeSqlNameUtils: SnowflakeSqlNameUtils
-
-    @BeforeEach
-    fun setUp() {
-        snowflakeConfiguration = mockk(relaxed = true)
-        snowflakeSqlNameUtils =
-            SnowflakeSqlNameUtils(snowflakeConfiguration = snowflakeConfiguration)
-    }
-
-    @Test
-    fun testFullyQualifiedName() {
-        val databaseName = "test-database"
-        val namespace = "test-namespace"
-        val name = "test=name"
-        val tableName = TableName(namespace = namespace, name = name)
-        every { snowflakeConfiguration.database } returns databaseName
-
-        val expectedName =
-            snowflakeSqlNameUtils.combineParts(
-                listOf(
-                    databaseName.toSnowflakeCompatibleName(),
-                    tableName.namespace,
-                    tableName.name
-                )
-            )
-        val fullyQualifiedName = snowflakeSqlNameUtils.fullyQualifiedName(tableName)
-        assertEquals(expectedName, fullyQualifiedName)
-    }
-
-    @Test
-    fun testFullyQualifiedNamespace() {
-        val databaseName = "test-database"
-        val namespace = "test-namespace"
-        every { snowflakeConfiguration.database } returns databaseName
-
-        val fullyQualifiedNamespace = snowflakeSqlNameUtils.fullyQualifiedNamespace(namespace)
-        assertEquals("\"TEST-DATABASE\".\"test-namespace\"", fullyQualifiedNamespace)
-    }
-
-    @Test
-    fun testFullyQualifiedStageName() {
-        val databaseName = "test-database"
-        val namespace = "test-namespace"
-        val name = "test=name"
-        val tableName = TableName(namespace = namespace, name = name)
-        every { snowflakeConfiguration.database } returns databaseName
-
-        val expectedName =
-            snowflakeSqlNameUtils.combineParts(
-                listOf(
-                    databaseName.toSnowflakeCompatibleName(),
-                    namespace,
-                    "$STAGE_NAME_PREFIX$name"
-                )
-            )
-        val fullyQualifiedName = snowflakeSqlNameUtils.fullyQualifiedStageName(tableName)
-        assertEquals(expectedName, fullyQualifiedName)
-    }
-
-    @Test
-    fun testFullyQualifiedStageNameWithEscape() {
-        val databaseName = "test-database"
-        val namespace = "test-namespace"
-        val name = "test=\"\"\'name"
-        val tableName = TableName(namespace = namespace, name = name)
-        every { snowflakeConfiguration.database } returns databaseName
-
-        val expectedName =
-            snowflakeSqlNameUtils.combineParts(
-                listOf(
-                    databaseName.toSnowflakeCompatibleName(),
-                    namespace,
-                    "$STAGE_NAME_PREFIX${sqlEscape(name)}"
-                )
-            )
-        val fullyQualifiedName = snowflakeSqlNameUtils.fullyQualifiedStageName(tableName, true)
-        assertEquals(expectedName, fullyQualifiedName)
-    }
-}

airbyte-integrations/connectors/destination-snowflake/src/test/kotlin/io/airbyte/integrations/destination/snowflake/write/SnowflakeWriterTest.kt

@@ -6,20 +6,21 @@ package io.airbyte.integrations.destination.snowflake.write
 
 import io.airbyte.cdk.SystemErrorException
 import io.airbyte.cdk.load.command.Append
+import io.airbyte.cdk.load.command.DestinationCatalog
 import io.airbyte.cdk.load.command.DestinationStream
-import io.airbyte.cdk.load.orchestration.db.DatabaseInitialStatusGatherer
-import io.airbyte.cdk.load.orchestration.db.TableNames
-import io.airbyte.cdk.load.orchestration.db.TempTableNameGenerator
-import io.airbyte.cdk.load.orchestration.db.direct_load_table.DirectLoadInitialStatus
-import io.airbyte.cdk.load.orchestration.db.direct_load_table.DirectLoadTableAppendStreamLoader
-import io.airbyte.cdk.load.orchestration.db.direct_load_table.DirectLoadTableAppendTruncateStreamLoader
-import io.airbyte.cdk.load.orchestration.db.direct_load_table.DirectLoadTableStatus
-import io.airbyte.cdk.load.orchestration.db.legacy_typing_deduping.TableCatalog
-import io.airbyte.cdk.load.orchestration.db.legacy_typing_deduping.TableNameInfo
-import io.airbyte.cdk.load.table.ColumnNameMapping
-import io.airbyte.cdk.load.table.TableName
+import io.airbyte.cdk.load.schema.model.ColumnSchema
+import io.airbyte.cdk.load.schema.model.StreamTableSchema
+import io.airbyte.cdk.load.schema.model.TableName
+import io.airbyte.cdk.load.schema.model.TableNames
+import io.airbyte.cdk.load.table.DatabaseInitialStatusGatherer
+import io.airbyte.cdk.load.table.TempTableNameGenerator
+import io.airbyte.cdk.load.table.directload.DirectLoadInitialStatus
+import io.airbyte.cdk.load.table.directload.DirectLoadTableAppendStreamLoader
+import io.airbyte.cdk.load.table.directload.DirectLoadTableAppendTruncateStreamLoader
+import io.airbyte.cdk.load.table.directload.DirectLoadTableExecutionConfig
+import io.airbyte.cdk.load.table.directload.DirectLoadTableStatus
+import io.airbyte.cdk.load.write.StreamStateStore
 import io.airbyte.integrations.destination.snowflake.client.SnowflakeAirbyteClient
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
 import io.mockk.coEvery
 import io.mockk.coVerify
 import io.mockk.every
@@ -34,55 +35,93 @@ internal class SnowflakeWriterTest {
     @Test
     fun testSetup() {
         val tableName = TableName(namespace = "test-namespace", name = "test-name")
-        val tableNames = TableNames(rawTableName = null, finalTableName = tableName)
-        val stream = mockk<DestinationStream>()
-        val tableInfo =
-            TableNameInfo(
-                tableNames = tableNames,
-                columnNameMapping = ColumnNameMapping(emptyMap())
-            )
-        val catalog = TableCatalog(mapOf(stream to tableInfo))
+        val tempTableName = TableName(namespace = "test-namespace", name = "test-name-temp")
+        val tableNames = TableNames(finalTableName = tableName, tempTableName = tempTableName)
+        val stream =
+            mockk<DestinationStream> {
+                every { tableSchema } returns
+                    StreamTableSchema(
+                        tableNames = tableNames,
+                        columnSchema =
+                            ColumnSchema(
+                                inputToFinalColumnNames = emptyMap(),
+                                finalSchema = emptyMap(),
+                                inputSchema = emptyMap()
+                            ),
+                        importType = Append
+                    )
+                every { mappedDescriptor } returns
+                    DestinationStream.Descriptor(
+                        namespace = tableName.namespace,
+                        name = tableName.name
+                    )
+                every { importType } returns Append
+            }
+        val catalog = DestinationCatalog(listOf(stream))
         val snowflakeClient = mockk<SnowflakeAirbyteClient>(relaxed = true)
         val stateGatherer =
             mockk<DatabaseInitialStatusGatherer<DirectLoadInitialStatus>> {
-                coEvery { gatherInitialStatus(catalog) } returns emptyMap()
+                coEvery { gatherInitialStatus() } returns
+                    mapOf(
+                        stream to
+                            DirectLoadInitialStatus(
+                                realTable = DirectLoadTableStatus(false),
+                                tempTable = null
+                            )
+                    )
             }
+        val streamStateStore = mockk<StreamStateStore<DirectLoadTableExecutionConfig>>()
         val writer =
             SnowflakeWriter(
-                names = catalog,
+                catalog = catalog,
                 stateGatherer = stateGatherer,
-                streamStateStore = mockk(),
+                streamStateStore = streamStateStore,
                 snowflakeClient = snowflakeClient,
                 tempTableNameGenerator = mockk(),
-                snowflakeConfiguration = mockk(relaxed = true),
+                snowflakeConfiguration =
+                    mockk(relaxed = true) {
+                        every { internalTableSchema } returns "internal_schema"
+                    },
             )
 
         runBlocking { writer.setup() }
 
         coVerify(exactly = 1) { snowflakeClient.createNamespace(tableName.namespace) }
-        coVerify(exactly = 1) { stateGatherer.gatherInitialStatus(catalog) }
+        coVerify(exactly = 1) { stateGatherer.gatherInitialStatus() }
     }
 
     @Test
     fun testCreateStreamLoaderFirstGeneration() {
         val tableName = TableName(namespace = "test-namespace", name = "test-name")
-        val tableNames = TableNames(rawTableName = null, finalTableName = tableName)
+        val tempTableName = TableName(namespace = "test-namespace", name = "test-name-temp")
+        val tableNames = TableNames(finalTableName = tableName, tempTableName = tempTableName)
         val stream =
             mockk<DestinationStream> {
                 every { minimumGenerationId } returns 0L
                 every { generationId } returns 0L
                 every { importType } returns Append
+                every { tableSchema } returns
+                    StreamTableSchema(
+                        tableNames = tableNames,
+                        columnSchema =
+                            ColumnSchema(
+                                inputToFinalColumnNames = emptyMap(),
+                                finalSchema = emptyMap(),
+                                inputSchema = emptyMap()
+                            ),
+                        importType = Append
+                    )
+                every { mappedDescriptor } returns
+                    DestinationStream.Descriptor(
+                        namespace = tableName.namespace,
+                        name = tableName.name
+                    )
             }
-        val tableInfo =
-            TableNameInfo(
-                tableNames = tableNames,
-                columnNameMapping = ColumnNameMapping(emptyMap())
-            )
-        val catalog = TableCatalog(mapOf(stream to tableInfo))
+        val catalog = DestinationCatalog(listOf(stream))
         val snowflakeClient = mockk<SnowflakeAirbyteClient>(relaxed = true)
         val stateGatherer =
             mockk<DatabaseInitialStatusGatherer<DirectLoadInitialStatus>> {
-                coEvery { gatherInitialStatus(catalog) } returns
+                coEvery { gatherInitialStatus() } returns
                     mapOf(
                         stream to
                             DirectLoadInitialStatus(
@@ -93,14 +132,18 @@ internal class SnowflakeWriterTest {
             }
         val tempTableNameGenerator =
             mockk<TempTableNameGenerator> { every { generate(any()) } answers { firstArg() } }
+        val streamStateStore = mockk<StreamStateStore<DirectLoadTableExecutionConfig>>()
         val writer =
             SnowflakeWriter(
-                names = catalog,
+                catalog = catalog,
                 stateGatherer = stateGatherer,
-                streamStateStore = mockk(),
+                streamStateStore = streamStateStore,
                 snowflakeClient = snowflakeClient,
                 tempTableNameGenerator = tempTableNameGenerator,
-                snowflakeConfiguration = mockk(relaxed = true),
+                snowflakeConfiguration =
+                    mockk(relaxed = true) {
+                        every { internalTableSchema } returns "internal_schema"
+                    },
             )
 
         runBlocking {
@@ -113,23 +156,35 @@ internal class SnowflakeWriterTest {
     @Test
     fun testCreateStreamLoaderNotFirstGeneration() {
         val tableName = TableName(namespace = "test-namespace", name = "test-name")
-        val tableNames = TableNames(rawTableName = null, finalTableName = tableName)
+        val tempTableName = TableName(namespace = "test-namespace", name = "test-name-temp")
+        val tableNames = TableNames(finalTableName = tableName, tempTableName = tempTableName)
         val stream =
             mockk<DestinationStream> {
                 every { minimumGenerationId } returns 1L
                 every { generationId } returns 1L
                 every { importType } returns Append
+                every { tableSchema } returns
+                    StreamTableSchema(
+                        tableNames = tableNames,
+                        columnSchema =
+                            ColumnSchema(
+                                inputToFinalColumnNames = emptyMap(),
+                                finalSchema = emptyMap(),
+                                inputSchema = emptyMap()
+                            ),
+                        importType = Append
+                    )
+                every { mappedDescriptor } returns
+                    DestinationStream.Descriptor(
+                        namespace = tableName.namespace,
+                        name = tableName.name
+                    )
             }
-        val tableInfo =
-            TableNameInfo(
-                tableNames = tableNames,
-                columnNameMapping = ColumnNameMapping(emptyMap())
-            )
-        val catalog = TableCatalog(mapOf(stream to tableInfo))
+        val catalog = DestinationCatalog(listOf(stream))
         val snowflakeClient = mockk<SnowflakeAirbyteClient>(relaxed = true)
         val stateGatherer =
             mockk<DatabaseInitialStatusGatherer<DirectLoadInitialStatus>> {
-                coEvery { gatherInitialStatus(catalog) } returns
+                coEvery { gatherInitialStatus() } returns
                     mapOf(
                         stream to
                             DirectLoadInitialStatus(
@@ -140,14 +195,18 @@ internal class SnowflakeWriterTest {
             }
         val tempTableNameGenerator =
             mockk<TempTableNameGenerator> { every { generate(any()) } answers { firstArg() } }
+        val streamStateStore = mockk<StreamStateStore<DirectLoadTableExecutionConfig>>()
         val writer =
             SnowflakeWriter(
-                names = catalog,
+                catalog = catalog,
                 stateGatherer = stateGatherer,
-                streamStateStore = mockk(),
+                streamStateStore = streamStateStore,
                 snowflakeClient = snowflakeClient,
                 tempTableNameGenerator = tempTableNameGenerator,
-                snowflakeConfiguration = mockk(relaxed = true),
+                snowflakeConfiguration =
+                    mockk(relaxed = true) {
+                        every { internalTableSchema } returns "internal_schema"
+                    },
             )
 
         runBlocking {
@@ -160,22 +219,35 @@ internal class SnowflakeWriterTest {
     @Test
     fun testCreateStreamLoaderHybrid() {
         val tableName = TableName(namespace = "test-namespace", name = "test-name")
-        val tableNames = TableNames(rawTableName = null, finalTableName = tableName)
+        val tempTableName = TableName(namespace = "test-namespace", name = "test-name-temp")
+        val tableNames = TableNames(finalTableName = tableName, tempTableName = tempTableName)
         val stream =
             mockk<DestinationStream> {
                 every { minimumGenerationId } returns 1L
                 every { generationId } returns 2L
+                every { importType } returns Append
+                every { tableSchema } returns
+                    StreamTableSchema(
+                        tableNames = tableNames,
+                        columnSchema =
+                            ColumnSchema(
+                                inputToFinalColumnNames = emptyMap(),
+                                finalSchema = emptyMap(),
+                                inputSchema = emptyMap()
+                            ),
+                        importType = Append
+                    )
+                every { mappedDescriptor } returns
+                    DestinationStream.Descriptor(
+                        namespace = tableName.namespace,
+                        name = tableName.name
+                    )
             }
-        val tableInfo =
-            TableNameInfo(
-                tableNames = tableNames,
-                columnNameMapping = ColumnNameMapping(emptyMap())
-            )
-        val catalog = TableCatalog(mapOf(stream to tableInfo))
+        val catalog = DestinationCatalog(listOf(stream))
         val snowflakeClient = mockk<SnowflakeAirbyteClient>(relaxed = true)
         val stateGatherer =
             mockk<DatabaseInitialStatusGatherer<DirectLoadInitialStatus>> {
-                coEvery { gatherInitialStatus(catalog) } returns
+                coEvery { gatherInitialStatus() } returns
                     mapOf(
                         stream to
                             DirectLoadInitialStatus(
@@ -186,14 +258,18 @@ internal class SnowflakeWriterTest {
             }
         val tempTableNameGenerator =
             mockk<TempTableNameGenerator> { every { generate(any()) } answers { firstArg() } }
+        val streamStateStore = mockk<StreamStateStore<DirectLoadTableExecutionConfig>>()
         val writer =
             SnowflakeWriter(
-                names = catalog,
+                catalog = catalog,
                 stateGatherer = stateGatherer,
-                streamStateStore = mockk(),
+                streamStateStore = streamStateStore,
                 snowflakeClient = snowflakeClient,
                 tempTableNameGenerator = tempTableNameGenerator,
-                snowflakeConfiguration = mockk(relaxed = true),
+                snowflakeConfiguration =
+                    mockk(relaxed = true) {
+                        every { internalTableSchema } returns "internal_schema"
+                    },
             )
 
         runBlocking {
@@ -203,169 +279,126 @@ internal class SnowflakeWriterTest {
     }
 
     @Test
-    fun testSetupWithNamespaceCreationFailure() {
-        val tableName = TableName(namespace = "test-namespace", name = "test-name")
-        val tableNames = TableNames(rawTableName = null, finalTableName = tableName)
-        val stream = mockk<DestinationStream>()
-        val tableInfo =
-            TableNameInfo(
-                tableNames = tableNames,
-                columnNameMapping = ColumnNameMapping(emptyMap())
-            )
-        val catalog = TableCatalog(mapOf(stream to tableInfo))
-        val snowflakeClient = mockk<SnowflakeAirbyteClient>()
-        val stateGatherer = mockk<DatabaseInitialStatusGatherer<DirectLoadInitialStatus>>()
-        val writer =
-            SnowflakeWriter(
-                names = catalog,
-                stateGatherer = stateGatherer,
-                streamStateStore = mockk(),
-                snowflakeClient = snowflakeClient,
-                tempTableNameGenerator = mockk(),
-                snowflakeConfiguration = mockk(),
-            )
-
-        // Simulate network failure during namespace creation
-        coEvery {
-            snowflakeClient.createNamespace(tableName.namespace.toSnowflakeCompatibleName())
-        } throws RuntimeException("Network connection failed")
-
-        assertThrows(RuntimeException::class.java) { runBlocking { writer.setup() } }
-    }
-
-    @Test
-    fun testSetupWithInitialStatusGatheringFailure() {
-        val tableName = TableName(namespace = "test-namespace", name = "test-name")
-        val tableNames = TableNames(rawTableName = null, finalTableName = tableName)
-        val stream = mockk<DestinationStream>()
-        val tableInfo =
-            TableNameInfo(
-                tableNames = tableNames,
-                columnNameMapping = ColumnNameMapping(emptyMap())
-            )
-        val catalog = TableCatalog(mapOf(stream to tableInfo))
-        val snowflakeClient = mockk<SnowflakeAirbyteClient>(relaxed = true)
-        val stateGatherer = mockk<DatabaseInitialStatusGatherer<DirectLoadInitialStatus>>()
-        val writer =
-            SnowflakeWriter(
-                names = catalog,
-                stateGatherer = stateGatherer,
-                streamStateStore = mockk(),
-                snowflakeClient = snowflakeClient,
-                tempTableNameGenerator = mockk(),
-                snowflakeConfiguration = mockk(),
-            )
-
-        // Simulate failure while gathering initial status
-        coEvery { stateGatherer.gatherInitialStatus(catalog) } throws
-            RuntimeException("Failed to query table status")
-
-        assertThrows(RuntimeException::class.java) { runBlocking { writer.setup() } }
-
-        // Verify namespace creation was still attempted
-        coVerify(exactly = 1) { snowflakeClient.createNamespace(tableName.namespace) }
-    }
-
-    @Test
-    fun testCreateStreamLoaderWithMissingInitialStatus() {
-        val tableName = TableName(namespace = "test-namespace", name = "test-name")
-        val tableNames = TableNames(rawTableName = null, finalTableName = tableName)
+    fun testCreateStreamLoaderNamespaceLegacy() {
+        val namespace = "test-namespace"
+        val name = "test-name"
+        val tableName = TableName(namespace = namespace, name = name)
+        val tempTableName = TableName(namespace = namespace, name = "${name}-temp")
+        val tableNames = TableNames(finalTableName = tableName, tempTableName = tempTableName)
         val stream =
             mockk<DestinationStream> {
                 every { minimumGenerationId } returns 0L
                 every { generationId } returns 0L
+                every { importType } returns Append
+                every { tableSchema } returns
+                    StreamTableSchema(
+                        tableNames = tableNames,
+                        columnSchema =
+                            ColumnSchema(
+                                inputToFinalColumnNames = emptyMap(),
+                                finalSchema = emptyMap(),
+                                inputSchema = emptyMap()
+                            ),
+                        importType = Append
+                    )
+                every { mappedDescriptor } returns
+                    DestinationStream.Descriptor(
+                        namespace = tableName.namespace,
+                        name = tableName.name
+                    )
             }
-        val missingStream =
-            mockk<DestinationStream> {
-                every { minimumGenerationId } returns 0L
-                every { generationId } returns 0L
-            }
-        val tableInfo =
-            TableNameInfo(
-                tableNames = tableNames,
-                columnNameMapping = ColumnNameMapping(emptyMap())
-            )
-        val catalog = TableCatalog(mapOf(stream to tableInfo))
+        val catalog = DestinationCatalog(listOf(stream))
         val snowflakeClient = mockk<SnowflakeAirbyteClient>(relaxed = true)
         val stateGatherer =
             mockk<DatabaseInitialStatusGatherer<DirectLoadInitialStatus>> {
-                coEvery { gatherInitialStatus(catalog) } returns
+                coEvery { gatherInitialStatus() } returns
                     mapOf(
                         stream to
                             DirectLoadInitialStatus(
                                 realTable = DirectLoadTableStatus(false),
-                                tempTable = null,
+                                tempTable = null
                             )
                     )
             }
+        val streamStateStore = mockk<StreamStateStore<DirectLoadTableExecutionConfig>>()
         val writer =
             SnowflakeWriter(
-                names = catalog,
+                catalog = catalog,
                 stateGatherer = stateGatherer,
-                streamStateStore = mockk(),
+                streamStateStore = streamStateStore,
                 snowflakeClient = snowflakeClient,
                 tempTableNameGenerator = mockk(),
-                snowflakeConfiguration = mockk(relaxed = true),
+                snowflakeConfiguration =
+                    mockk(relaxed = true) {
+                        every { legacyRawTablesOnly } returns true
+                        every { internalTableSchema } returns "internal_schema"
+                    },
             )
 
-        runBlocking {
-            writer.setup()
-            // Try to create loader for a stream that wasn't in initial status
-            assertThrows(NullPointerException::class.java) {
-                writer.createStreamLoader(missingStream)
+        runBlocking { writer.setup() }
+
+        coVerify(exactly = 1) { snowflakeClient.createNamespace(tableName.namespace) }
+    }
+
+    @Test
+    fun testCreateStreamLoaderNamespaceNonLegacy() {
+        val namespace = "test-namespace"
+        val name = "test-name"
+        val tableName = TableName(namespace = namespace, name = name)
+        val tempTableName = TableName(namespace = namespace, name = "${name}-temp")
+        val tableNames = TableNames(finalTableName = tableName, tempTableName = tempTableName)
+        val stream =
+            mockk<DestinationStream> {
+                every { minimumGenerationId } returns 0L
+                every { generationId } returns 0L
+                every { importType } returns Append
+                every { tableSchema } returns
+                    StreamTableSchema(
+                        tableNames = tableNames,
+                        columnSchema =
+                            ColumnSchema(
+                                inputToFinalColumnNames = emptyMap(),
+                                finalSchema = emptyMap(),
+                                inputSchema = emptyMap()
+                            ),
+                        importType = Append
+                    )
+                every { mappedDescriptor } returns
+                    DestinationStream.Descriptor(
+                        namespace = tableName.namespace,
+                        name = tableName.name
+                    )
             }
-        }
-    }
-
-    @Test
-    fun testCreateStreamLoaderWithNullFinalTableName() {
-        // TableNames constructor throws IllegalStateException when both names are null
-        assertThrows(IllegalStateException::class.java) {
-            TableNames(rawTableName = null, finalTableName = null)
-        }
-    }
-
-    @Test
-    fun testSetupWithMultipleNamespaceFailuresPartial() {
-        val tableName1 = TableName(namespace = "namespace1", name = "table1")
-        val tableName2 = TableName(namespace = "namespace2", name = "table2")
-        val tableNames1 = TableNames(rawTableName = null, finalTableName = tableName1)
-        val tableNames2 = TableNames(rawTableName = null, finalTableName = tableName2)
-        val stream1 = mockk<DestinationStream>()
-        val stream2 = mockk<DestinationStream>()
-        val tableInfo1 =
-            TableNameInfo(
-                tableNames = tableNames1,
-                columnNameMapping = ColumnNameMapping(emptyMap())
-            )
-        val tableInfo2 =
-            TableNameInfo(
-                tableNames = tableNames2,
-                columnNameMapping = ColumnNameMapping(emptyMap())
-            )
-        val catalog = TableCatalog(mapOf(stream1 to tableInfo1, stream2 to tableInfo2))
-        val snowflakeClient = mockk<SnowflakeAirbyteClient>()
-        val stateGatherer = mockk<DatabaseInitialStatusGatherer<DirectLoadInitialStatus>>()
+        val catalog = DestinationCatalog(listOf(stream))
+        val snowflakeClient = mockk<SnowflakeAirbyteClient>(relaxed = true)
+        val stateGatherer =
+            mockk<DatabaseInitialStatusGatherer<DirectLoadInitialStatus>> {
+                coEvery { gatherInitialStatus() } returns
+                    mapOf(
+                        stream to
+                            DirectLoadInitialStatus(
+                                realTable = DirectLoadTableStatus(false),
+                                tempTable = null
+                            )
+                    )
+            }
+        val streamStateStore = mockk<StreamStateStore<DirectLoadTableExecutionConfig>>()
         val writer =
             SnowflakeWriter(
-                names = catalog,
+                catalog = catalog,
                 stateGatherer = stateGatherer,
-                streamStateStore = mockk(),
+                streamStateStore = streamStateStore,
                 snowflakeClient = snowflakeClient,
                 tempTableNameGenerator = mockk(),
-                snowflakeConfiguration = mockk(),
+                snowflakeConfiguration =
+                    mockk(relaxed = true) {
+                        every { legacyRawTablesOnly } returns false
+                        every { internalTableSchema } returns "internal_schema"
+                    },
             )
 
-        // First namespace succeeds, second fails (namespaces are uppercased by
-        // toSnowflakeCompatibleName)
-        coEvery { snowflakeClient.createNamespace("namespace1") } returns Unit
-        coEvery { snowflakeClient.createNamespace("namespace2") } throws
-            RuntimeException("Connection timeout")
+        runBlocking { writer.setup() }
 
-        assertThrows(RuntimeException::class.java) { runBlocking { writer.setup() } }
-
-        // Verify both namespace creations were attempted
-        coVerify(exactly = 1) { snowflakeClient.createNamespace("namespace1") }
-        coVerify(exactly = 1) { snowflakeClient.createNamespace("namespace2") }
+        coVerify(exactly = 1) { snowflakeClient.createNamespace(namespace) }
     }
 }

airbyte-integrations/connectors/destination-snowflake/src/test/kotlin/io/airbyte/integrations/destination/snowflake/write/load/SnowflakeInsertBufferTest.kt

@@ -4,218 +4,220 @@
 
 package io.airbyte.integrations.destination.snowflake.write.load
 
+import io.airbyte.cdk.load.component.ColumnType
 import io.airbyte.cdk.load.data.AirbyteValue
+import io.airbyte.cdk.load.data.FieldType
 import io.airbyte.cdk.load.data.IntegerValue
 import io.airbyte.cdk.load.data.NullValue
+import io.airbyte.cdk.load.data.StringType
 import io.airbyte.cdk.load.data.StringValue
 import io.airbyte.cdk.load.message.Meta
-import io.airbyte.cdk.load.table.TableName
+import io.airbyte.cdk.load.schema.model.ColumnSchema
+import io.airbyte.cdk.load.schema.model.TableName
 import io.airbyte.integrations.destination.snowflake.client.SnowflakeAirbyteClient
+import io.airbyte.integrations.destination.snowflake.schema.SnowflakeColumnManager
 import io.airbyte.integrations.destination.snowflake.spec.SnowflakeConfiguration
-import io.airbyte.integrations.destination.snowflake.sql.SnowflakeColumnUtils
 import io.mockk.coVerify
 import io.mockk.every
 import io.mockk.mockk
 import java.io.BufferedReader
-import java.io.File
 import java.io.InputStreamReader
 import java.util.zip.GZIPInputStream
 import kotlin.io.path.exists
 import kotlinx.coroutines.runBlocking
 import org.junit.jupiter.api.Assertions.assertEquals
+import org.junit.jupiter.api.Assertions.assertNotNull
 import org.junit.jupiter.api.BeforeEach
 import org.junit.jupiter.api.Test
 
 internal class SnowflakeInsertBufferTest {
 
     private lateinit var snowflakeConfiguration: SnowflakeConfiguration
-    private lateinit var snowflakeColumnUtils: SnowflakeColumnUtils
+    private lateinit var columnManager: SnowflakeColumnManager
+    private lateinit var columnSchema: ColumnSchema
+    private lateinit var snowflakeRecordFormatter: SnowflakeRecordFormatter
 
     @BeforeEach
     fun setUp() {
         snowflakeConfiguration = mockk(relaxed = true)
-        snowflakeColumnUtils = mockk(relaxed = true)
+        snowflakeRecordFormatter = SnowflakeSchemaRecordFormatter()
+        columnManager =
+            mockk(relaxed = true) {
+                every { getMetaColumns() } returns
+                    linkedMapOf(
+                        "_AIRBYTE_RAW_ID" to ColumnType("VARCHAR", false),
+                        "_AIRBYTE_EXTRACTED_AT" to ColumnType("TIMESTAMP_TZ", false),
+                        "_AIRBYTE_META" to ColumnType("VARIANT", false),
+                        "_AIRBYTE_GENERATION_ID" to ColumnType("NUMBER", true)
+                    )
+                every { getTableColumnNames(any()) } returns
+                    listOf(
+                        "_AIRBYTE_RAW_ID",
+                        "_AIRBYTE_EXTRACTED_AT",
+                        "_AIRBYTE_META",
+                        "_AIRBYTE_GENERATION_ID",
+                        "columnName"
+                    )
+            }
     }
 
     @Test
     fun testAccumulate() {
-        val tableName = mockk<TableName>(relaxed = true)
+        val tableName = TableName(namespace = "test", name = "table")
         val column = "columnName"
-        val columns = linkedMapOf(column to "NUMBER(38,0)")
+        columnSchema =
+            ColumnSchema(
+                inputToFinalColumnNames = mapOf(column to column.uppercase()),
+                finalSchema = mapOf(column.uppercase() to ColumnType("NUMBER", true)),
+                inputSchema = mapOf(column to FieldType(StringType, nullable = true))
+            )
         val snowflakeAirbyteClient = mockk<SnowflakeAirbyteClient>(relaxed = true)
         val record = createRecord(column)
         val buffer =
             SnowflakeInsertBuffer(
                 tableName = tableName,
-                columns = columns,
                 snowflakeClient = snowflakeAirbyteClient,
                 snowflakeConfiguration = snowflakeConfiguration,
-                flushLimit = 1,
-                snowflakeColumnUtils = snowflakeColumnUtils,
+                columnSchema = columnSchema,
+                columnManager = columnManager,
+                snowflakeRecordFormatter = snowflakeRecordFormatter,
             )
-
-        buffer.accumulate(record)
-
-        assertEquals(true, buffer.csvFilePath?.exists())
+        assertEquals(0, buffer.recordCount)
+        runBlocking { buffer.accumulate(record) }
         assertEquals(1, buffer.recordCount)
     }
 
     @Test
-    fun testAccumulateRaw() {
-        val tableName = mockk<TableName>(relaxed = true)
+    fun testFlushToStaging() {
+        val tableName = TableName(namespace = "test", name = "table")
         val column = "columnName"
-        val columns = linkedMapOf(column to "NUMBER(38,0)")
+        columnSchema =
+            ColumnSchema(
+                inputToFinalColumnNames = mapOf(column to column.uppercase()),
+                finalSchema = mapOf(column.uppercase() to ColumnType("NUMBER", true)),
+                inputSchema = mapOf(column to FieldType(StringType, nullable = true))
+            )
         val snowflakeAirbyteClient = mockk<SnowflakeAirbyteClient>(relaxed = true)
         val record = createRecord(column)
         val buffer =
             SnowflakeInsertBuffer(
                 tableName = tableName,
-                columns = columns,
                 snowflakeClient = snowflakeAirbyteClient,
                 snowflakeConfiguration = snowflakeConfiguration,
+                columnSchema = columnSchema,
+                columnManager = columnManager,
+                snowflakeRecordFormatter = snowflakeRecordFormatter,
                 flushLimit = 1,
-                snowflakeColumnUtils = snowflakeColumnUtils,
             )
-
-        every { snowflakeConfiguration.legacyRawTablesOnly } returns true
-
-        buffer.accumulate(record)
-
-        assertEquals(true, buffer.csvFilePath?.exists())
-        assertEquals(1, buffer.recordCount)
-    }
-
-    @Test
-    fun testFlush() {
-        val tableName = mockk<TableName>(relaxed = true)
-        val column = "columnName"
-        val columns = linkedMapOf(column to "NUMBER(38,0)")
-        val snowflakeAirbyteClient = mockk<SnowflakeAirbyteClient>(relaxed = true)
-        val record = createRecord(column)
-        val buffer =
-            SnowflakeInsertBuffer(
-                tableName = tableName,
-                columns = columns,
-                snowflakeClient = snowflakeAirbyteClient,
-                snowflakeConfiguration = snowflakeConfiguration,
-                flushLimit = 1,
-                snowflakeColumnUtils = snowflakeColumnUtils,
+        val expectedColumnNames =
+            listOf(
+                "_AIRBYTE_RAW_ID",
+                "_AIRBYTE_EXTRACTED_AT",
+                "_AIRBYTE_META",
+                "_AIRBYTE_GENERATION_ID",
+                "columnName"
             )
-
         runBlocking {
             buffer.accumulate(record)
             buffer.flush()
-        }
-
-        coVerify(exactly = 1) { snowflakeAirbyteClient.putInStage(tableName, any()) }
-        coVerify(exactly = 1) {
-            snowflakeAirbyteClient.copyFromStage(
-                tableName,
-                match { it.endsWith("$CSV_FILE_EXTENSION$FILE_SUFFIX") }
-            )
+            coVerify(exactly = 1) { snowflakeAirbyteClient.putInStage(tableName, any()) }
+            coVerify(exactly = 1) {
+                snowflakeAirbyteClient.copyFromStage(tableName, any(), expectedColumnNames)
+            }
         }
     }
 
     @Test
-    fun testFlushRaw() {
-        val tableName = mockk<TableName>(relaxed = true)
+    fun testFlushToNoStaging() {
+        val tableName = TableName(namespace = "test", name = "table")
         val column = "columnName"
-        val columns = linkedMapOf(column to "NUMBER(38,0)")
+        columnSchema =
+            ColumnSchema(
+                inputToFinalColumnNames = mapOf(column to column.uppercase()),
+                finalSchema = mapOf(column.uppercase() to ColumnType("NUMBER", true)),
+                inputSchema = mapOf(column to FieldType(StringType, nullable = true))
+            )
+        val snowflakeAirbyteClient = mockk<SnowflakeAirbyteClient>(relaxed = true)
+        every { snowflakeConfiguration.legacyRawTablesOnly } returns true
+        val record = createRecord(column)
+        val buffer =
+            SnowflakeInsertBuffer(
+                tableName = tableName,
+                snowflakeClient = snowflakeAirbyteClient,
+                snowflakeConfiguration = snowflakeConfiguration,
+                columnSchema = columnSchema,
+                columnManager = columnManager,
+                snowflakeRecordFormatter = snowflakeRecordFormatter,
+                flushLimit = 1,
+            )
+        val expectedColumnNames =
+            listOf(
+                "_AIRBYTE_RAW_ID",
+                "_AIRBYTE_EXTRACTED_AT",
+                "_AIRBYTE_META",
+                "_AIRBYTE_GENERATION_ID",
+                "columnName"
+            )
+        runBlocking {
+            buffer.accumulate(record)
+            buffer.flush()
+            // In legacy raw mode, it still uses staging
+            coVerify(exactly = 1) { snowflakeAirbyteClient.putInStage(tableName, any()) }
+            coVerify(exactly = 1) {
+                snowflakeAirbyteClient.copyFromStage(tableName, any(), expectedColumnNames)
+            }
+        }
+    }
+
+    @Test
+    fun testFileCreation() {
+        val tableName = TableName(namespace = "test", name = "table")
+        val column = "columnName"
+        columnSchema =
+            ColumnSchema(
+                inputToFinalColumnNames = mapOf(column to column.uppercase()),
+                finalSchema = mapOf(column.uppercase() to ColumnType("NUMBER", true)),
+                inputSchema = mapOf(column to FieldType(StringType, nullable = true))
+            )
         val snowflakeAirbyteClient = mockk<SnowflakeAirbyteClient>(relaxed = true)
         val record = createRecord(column)
         val buffer =
             SnowflakeInsertBuffer(
                 tableName = tableName,
-                columns = columns,
                 snowflakeClient = snowflakeAirbyteClient,
                 snowflakeConfiguration = snowflakeConfiguration,
+                columnSchema = columnSchema,
+                columnManager = columnManager,
+                snowflakeRecordFormatter = snowflakeRecordFormatter,
                 flushLimit = 1,
-                snowflakeColumnUtils = snowflakeColumnUtils,
             )
-
-        every { snowflakeConfiguration.legacyRawTablesOnly } returns true
-
         runBlocking {
             buffer.accumulate(record)
-            buffer.flush()
-        }
-
-        coVerify(exactly = 1) { snowflakeAirbyteClient.putInStage(tableName, any()) }
-        coVerify(exactly = 1) {
-            snowflakeAirbyteClient.copyFromStage(
-                tableName,
-                match { it.endsWith("$CSV_FILE_EXTENSION$FILE_SUFFIX") }
-            )
-        }
-    }
-
-    @Test
-    fun testMissingFields() {
-        val tableName = mockk<TableName>(relaxed = true)
-        val snowflakeAirbyteClient = mockk<SnowflakeAirbyteClient>(relaxed = true)
-        val record = createRecord("COLUMN1")
-        val buffer =
-            SnowflakeInsertBuffer(
-                tableName = tableName,
-                columns = linkedMapOf("COLUMN1" to "NUMBER(38,0)", "COLUMN2" to "NUMBER(38,0)"),
-                snowflakeClient = snowflakeAirbyteClient,
-                snowflakeConfiguration = snowflakeConfiguration,
-                flushLimit = 1,
-                snowflakeColumnUtils = snowflakeColumnUtils,
-            )
-
-        runBlocking {
-            buffer.accumulate(record)
-            buffer.csvWriter?.flush()
+            // The csvFilePath is internal, we can access it for testing
+            val filepath = buffer.csvFilePath
+            assertNotNull(filepath)
+            val file = filepath!!.toFile()
+            assert(file.exists())
+            // Close the writer to ensure all data is flushed
             buffer.csvWriter?.close()
-            assertEquals(
-                "test-value$CSV_FIELD_SEPARATOR$CSV_LINE_DELIMITER",
-                readFromCsvFile(buffer.csvFilePath!!.toFile())
-            )
+            val lines = mutableListOf<String>()
+            GZIPInputStream(file.inputStream()).use { gzip ->
+                BufferedReader(InputStreamReader(gzip)).use { bufferedReader ->
+                    bufferedReader.forEachLine { line -> lines.add(line) }
+                }
+            }
+            assertEquals(1, lines.size)
+            file.delete()
         }
     }
 
-    @Test
-    fun testMissingFieldsRaw() {
-        val tableName = mockk<TableName>(relaxed = true)
-        val snowflakeAirbyteClient = mockk<SnowflakeAirbyteClient>(relaxed = true)
-        val record = createRecord("COLUMN1")
-        val buffer =
-            SnowflakeInsertBuffer(
-                tableName = tableName,
-                columns = linkedMapOf("COLUMN1" to "NUMBER(38,0)", "COLUMN2" to "NUMBER(38,0)"),
-                snowflakeClient = snowflakeAirbyteClient,
-                snowflakeConfiguration = snowflakeConfiguration,
-                flushLimit = 1,
-                snowflakeColumnUtils = snowflakeColumnUtils,
-            )
-
-        every { snowflakeConfiguration.legacyRawTablesOnly } returns true
-
-        runBlocking {
-            buffer.accumulate(record)
-            buffer.csvWriter?.flush()
-            buffer.csvWriter?.close()
-            assertEquals(
-                "test-value$CSV_FIELD_SEPARATOR$CSV_LINE_DELIMITER",
-                readFromCsvFile(buffer.csvFilePath!!.toFile())
-            )
-        }
-    }
-
-    private fun readFromCsvFile(file: File) =
-        GZIPInputStream(file.inputStream()).use { input ->
-            val reader = BufferedReader(InputStreamReader(input))
-            reader.readText()
-        }
-
-    private fun createRecord(columnName: String) =
-        mapOf(
-            columnName to AirbyteValue.from("test-value"),
-            Meta.COLUMN_NAME_AB_EXTRACTED_AT to IntegerValue(System.currentTimeMillis()),
-            Meta.COLUMN_NAME_AB_RAW_ID to StringValue("raw-id"),
-            Meta.COLUMN_NAME_AB_GENERATION_ID to IntegerValue(1223),
-            Meta.COLUMN_NAME_AB_META to StringValue("{\"changes\":[],\"syncId\":43}"),
-            "${columnName}Null" to NullValue
+    private fun createRecord(column: String): Map<String, AirbyteValue> {
+        return mapOf(
+            column to IntegerValue(value = 42),
+            Meta.COLUMN_NAME_AB_GENERATION_ID to NullValue,
+            Meta.COLUMN_NAME_AB_RAW_ID to StringValue("raw-id-1"),
+            Meta.COLUMN_NAME_AB_EXTRACTED_AT to IntegerValue(1234567890),
+            Meta.COLUMN_NAME_AB_META to StringValue("meta-data-foo"),
         )
+    }
 }

airbyte-integrations/connectors/destination-snowflake/src/test/kotlin/io/airbyte/integrations/destination/snowflake/write/load/SnowflakeRawRecordFormatterTest.kt

@@ -15,12 +15,8 @@ import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_LOADED_AT
 import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_META
 import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_RAW_ID
 import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_DATA
-import io.airbyte.integrations.destination.snowflake.sql.SnowflakeColumnUtils
-import io.mockk.every
-import io.mockk.mockk
-import kotlin.collections.plus
+import io.airbyte.cdk.load.schema.model.ColumnSchema
 import org.junit.jupiter.api.Assertions.assertEquals
-import org.junit.jupiter.api.BeforeEach
 import org.junit.jupiter.api.Test
 
 private val AIRBYTE_COLUMN_TYPES_MAP =
@@ -58,28 +54,16 @@ private fun createExpected(
 
 internal class SnowflakeRawRecordFormatterTest {
 
-    private lateinit var snowflakeColumnUtils: SnowflakeColumnUtils
-
-    @BeforeEach
-    fun setup() {
-        snowflakeColumnUtils = mockk {
-            every { getFormattedDefaultColumnNames(any()) } returns
-                AIRBYTE_COLUMN_TYPES_MAP.keys.toList()
-        }
-    }
-
     @Test
     fun testFormatting() {
         val columnName = "test-column-name"
         val columnValue = "test-column-value"
         val columns = AIRBYTE_COLUMN_TYPES_MAP
         val record = createRecord(columnName = columnName, columnValue = columnValue)
-        val formatter =
-            SnowflakeRawRecordFormatter(
-                columns = AIRBYTE_COLUMN_TYPES_MAP,
-                snowflakeColumnUtils = snowflakeColumnUtils
-            )
-        val formattedValue = formatter.format(record)
+        val formatter = SnowflakeRawRecordFormatter()
+        // RawRecordFormatter doesn't use columnSchema but still needs one per interface
+        val dummyColumnSchema = ColumnSchema(emptyMap(), emptyMap(), emptyMap())
+        val formattedValue = formatter.format(record, dummyColumnSchema)
         val expectedValue =
             createExpected(
                 record = record,
@@ -93,33 +77,28 @@ internal class SnowflakeRawRecordFormatterTest {
     fun testFormattingMigratedFromPreviousVersion() {
         val columnName = "test-column-name"
         val columnValue = "test-column-value"
-        val columnsMap =
-            linkedMapOf(
-                COLUMN_NAME_AB_EXTRACTED_AT to "TIMESTAMP_TZ(9)",
-                COLUMN_NAME_AB_LOADED_AT to "TIMESTAMP_TZ(9)",
-                COLUMN_NAME_AB_META to "VARIANT",
-                COLUMN_NAME_DATA to "VARIANT",
-                COLUMN_NAME_AB_RAW_ID to "VARCHAR(16777216)",
-                COLUMN_NAME_AB_GENERATION_ID to "NUMBER(38,0)",
-            )
         val record = createRecord(columnName = columnName, columnValue = columnValue)
-        val formatter =
-            SnowflakeRawRecordFormatter(
-                columns = columnsMap,
-                snowflakeColumnUtils = snowflakeColumnUtils
-            )
-        val formattedValue = formatter.format(record)
+        val formatter = SnowflakeRawRecordFormatter()
+        // RawRecordFormatter doesn't use columnSchema but still needs one per interface
+        val dummyColumnSchema = ColumnSchema(emptyMap(), emptyMap(), emptyMap())
+        val formattedValue = formatter.format(record, dummyColumnSchema)
+
+        // The formatter outputs in a fixed order regardless of input column order:
+        // 1. AB_RAW_ID
+        // 2. AB_EXTRACTED_AT
+        // 3. AB_META
+        // 4. AB_GENERATION_ID
+        // 5. AB_LOADED_AT
+        // 6. DATA (JSON with remaining columns)
         val expectedValue =
-            createExpected(
-                    record = record,
-                    columns = columnsMap,
-                    airbyteColumns = columnsMap.keys.toList(),
-                )
-                .toMutableList()
-        expectedValue.add(
-            columnsMap.keys.indexOf(COLUMN_NAME_DATA),
-            "{\"$columnName\":\"$columnValue\"}"
-        )
+            listOf(
+                record[COLUMN_NAME_AB_RAW_ID]!!.toCsvValue(),
+                record[COLUMN_NAME_AB_EXTRACTED_AT]!!.toCsvValue(),
+                record[COLUMN_NAME_AB_META]!!.toCsvValue(),
+                record[COLUMN_NAME_AB_GENERATION_ID]!!.toCsvValue(),
+                record[COLUMN_NAME_AB_LOADED_AT]!!.toCsvValue(),
+                "{\"$columnName\":\"$columnValue\"}"
+            )
         assertEquals(expectedValue, formattedValue)
     }
 }

airbyte-integrations/connectors/destination-snowflake/src/test/kotlin/io/airbyte/integrations/destination/snowflake/write/load/SnowflakeSchemaRecordFormatterTest.kt

@@ -4,66 +4,58 @@
 
 package io.airbyte.integrations.destination.snowflake.write.load
 
+import io.airbyte.cdk.load.component.ColumnType
 import io.airbyte.cdk.load.data.AirbyteValue
+import io.airbyte.cdk.load.data.FieldType
 import io.airbyte.cdk.load.data.IntegerValue
 import io.airbyte.cdk.load.data.NullValue
+import io.airbyte.cdk.load.data.StringType
 import io.airbyte.cdk.load.data.StringValue
 import io.airbyte.cdk.load.data.csv.toCsvValue
 import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_EXTRACTED_AT
 import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_GENERATION_ID
 import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_META
 import io.airbyte.cdk.load.message.Meta.Companion.COLUMN_NAME_AB_RAW_ID
-import io.airbyte.integrations.destination.snowflake.db.toSnowflakeCompatibleName
-import io.airbyte.integrations.destination.snowflake.sql.SnowflakeColumnUtils
-import io.mockk.every
-import io.mockk.mockk
-import java.util.AbstractMap
-import kotlin.collections.component1
-import kotlin.collections.component2
+import io.airbyte.cdk.load.schema.model.ColumnSchema
+import io.airbyte.integrations.destination.snowflake.schema.toSnowflakeCompatibleName
 import org.junit.jupiter.api.Assertions.assertEquals
-import org.junit.jupiter.api.BeforeEach
 import org.junit.jupiter.api.Test
 
-private val AIRBYTE_COLUMN_TYPES_MAP =
-    linkedMapOf(
-            COLUMN_NAME_AB_RAW_ID to "VARCHAR(16777216)",
-            COLUMN_NAME_AB_EXTRACTED_AT to "TIMESTAMP_TZ(9)",
-            COLUMN_NAME_AB_META to "VARIANT",
-            COLUMN_NAME_AB_GENERATION_ID to "NUMBER(38,0)",
-        )
-        .mapKeys { it.key.toSnowflakeCompatibleName() }
-
 internal class SnowflakeSchemaRecordFormatterTest {
 
-    private lateinit var snowflakeColumnUtils: SnowflakeColumnUtils
+    private fun createColumnSchema(userColumns: Map<String, String>): ColumnSchema {
+        val finalSchema = linkedMapOf<String, ColumnType>()
+        val inputToFinalColumnNames = mutableMapOf<String, String>()
+        val inputSchema = mutableMapOf<String, FieldType>()
 
-    @BeforeEach
-    fun setup() {
-        snowflakeColumnUtils = mockk {
-            every { getFormattedDefaultColumnNames(any()) } returns
-                AIRBYTE_COLUMN_TYPES_MAP.keys.toList()
+        // Add user columns
+        userColumns.forEach { (name, type) ->
+            val finalName = name.toSnowflakeCompatibleName()
+            finalSchema[finalName] = ColumnType(type, true)
+            inputToFinalColumnNames[name] = finalName
+            inputSchema[name] = FieldType(StringType, nullable = true)
         }
+
+        return ColumnSchema(
+            inputToFinalColumnNames = inputToFinalColumnNames,
+            finalSchema = finalSchema,
+            inputSchema = inputSchema
+        )
     }
 
     @Test
     fun testFormatting() {
         val columnName = "test-column-name"
         val columnValue = "test-column-value"
-        val columns =
-            (AIRBYTE_COLUMN_TYPES_MAP + linkedMapOf(columnName to "VARCHAR(16777216)")).mapKeys {
-                it.key.toSnowflakeCompatibleName()
-            }
+        val userColumns = mapOf(columnName to "VARCHAR(16777216)")
+        val columnSchema = createColumnSchema(userColumns)
         val record = createRecord(columnName, columnValue)
-        val formatter =
-            SnowflakeSchemaRecordFormatter(
-                columns = columns as LinkedHashMap<String, String>,
-                snowflakeColumnUtils = snowflakeColumnUtils
-            )
-        val formattedValue = formatter.format(record)
+        val formatter = SnowflakeSchemaRecordFormatter()
+        val formattedValue = formatter.format(record, columnSchema)
         val expectedValue =
             createExpected(
                 record = record,
-                columns = columns,
+                columnSchema = columnSchema,
             )
         assertEquals(expectedValue, formattedValue)
     }
@@ -72,21 +64,15 @@ internal class SnowflakeSchemaRecordFormatterTest {
     fun testFormattingVariant() {
         val columnName = "test-column-name"
         val columnValue = "{\"test\": \"test-value\"}"
-        val columns =
-            (AIRBYTE_COLUMN_TYPES_MAP + linkedMapOf(columnName to "VARIANT")).mapKeys {
-                it.key.toSnowflakeCompatibleName()
-            }
+        val userColumns = mapOf(columnName to "VARIANT")
+        val columnSchema = createColumnSchema(userColumns)
         val record = createRecord(columnName, columnValue)
-        val formatter =
-            SnowflakeSchemaRecordFormatter(
-                columns = columns as LinkedHashMap<String, String>,
-                snowflakeColumnUtils = snowflakeColumnUtils
-            )
-        val formattedValue = formatter.format(record)
+        val formatter = SnowflakeSchemaRecordFormatter()
+        val formattedValue = formatter.format(record, columnSchema)
         val expectedValue =
             createExpected(
                 record = record,
-                columns = columns,
+                columnSchema = columnSchema,
             )
         assertEquals(expectedValue, formattedValue)
     }
@@ -95,23 +81,16 @@ internal class SnowflakeSchemaRecordFormatterTest {
     fun testFormattingMissingColumn() {
         val columnName = "test-column-name"
         val columnValue = "test-column-value"
-        val columns =
-            AIRBYTE_COLUMN_TYPES_MAP +
-                linkedMapOf(
-                    columnName to "VARCHAR(16777216)",
-                    "missing-column" to "VARCHAR(16777216)"
-                )
+        val userColumns =
+            mapOf(columnName to "VARCHAR(16777216)", "missing-column" to "VARCHAR(16777216)")
+        val columnSchema = createColumnSchema(userColumns)
         val record = createRecord(columnName, columnValue)
-        val formatter =
-            SnowflakeSchemaRecordFormatter(
-                columns = columns as LinkedHashMap<String, String>,
-                snowflakeColumnUtils = snowflakeColumnUtils
-            )
-        val formattedValue = formatter.format(record)
+        val formatter = SnowflakeSchemaRecordFormatter()
+        val formattedValue = formatter.format(record, columnSchema)
         val expectedValue =
             createExpected(
                 record = record,
-                columns = columns,
+                columnSchema = columnSchema,
                 filterMissing = false,
             )
         assertEquals(expectedValue, formattedValue)
@@ -128,16 +107,37 @@ internal class SnowflakeSchemaRecordFormatterTest {
 
     private fun createExpected(
         record: Map<String, AirbyteValue>,
-        columns: Map<String, String>,
+        columnSchema: ColumnSchema,
         filterMissing: Boolean = true,
-    ) =
-        record.entries
-            .associate { entry -> entry.key.toSnowflakeCompatibleName() to entry.value }
-            .map { entry -> AbstractMap.SimpleEntry(entry.key, entry.value.toCsvValue()) }
-            .sortedBy { entry ->
-                if (columns.keys.indexOf(entry.key) > -1) columns.keys.indexOf(entry.key)
-                else Int.MAX_VALUE
+    ): List<Any> {
+        val columns = columnSchema.finalSchema.keys.toList()
+        val result = mutableListOf<Any>()
+
+        // Add meta columns first in the expected order
+        result.add(record[COLUMN_NAME_AB_RAW_ID]?.toCsvValue() ?: "")
+        result.add(record[COLUMN_NAME_AB_EXTRACTED_AT]?.toCsvValue() ?: "")
+        result.add(record[COLUMN_NAME_AB_META]?.toCsvValue() ?: "")
+        result.add(record[COLUMN_NAME_AB_GENERATION_ID]?.toCsvValue() ?: "")
+
+        // Add user columns
+        val userColumns =
+            columns.filterNot { col ->
+                listOf(
+                        COLUMN_NAME_AB_RAW_ID.toSnowflakeCompatibleName(),
+                        COLUMN_NAME_AB_EXTRACTED_AT.toSnowflakeCompatibleName(),
+                        COLUMN_NAME_AB_META.toSnowflakeCompatibleName(),
+                        COLUMN_NAME_AB_GENERATION_ID.toSnowflakeCompatibleName()
+                    )
+                    .contains(col)
             }
-            .filter { (k, _) -> if (filterMissing) columns.contains(k) else true }
-            .map { it.value }
+
+        userColumns.forEach { columnName ->
+            val value = record[columnName] ?: if (!filterMissing) NullValue else null
+            if (value != null || !filterMissing) {
+                result.add(value?.toCsvValue() ?: "")
+            }
+        }
+
+        return result
+    }
 }