Trino ROW (struct) columns were serialized as plain arrays, losing
all field name information. This converts NamedRowTuple values to
dicts using the field name metadata from the trino library, so ROW
fields appear as {"field": value} instead of [value].
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: Tsuneo Yoshioka <yoshiokatsuneo@gmail.com>
* Update Azure Data Explorer query runner to latest version
* Fix black issue
* downgrade azure-kusto-data to 4.6.3
* Freeze numpy to 1.24.4 because of 2.0.0 incompatibility
* Fix failing test
* Reformat test
This reverts commit bd115e7f5f, as
it turns out to be a useful security feature.
In order to remove this in a better way, we'll need to replace it
with something that provides equivalent functionality.
Since #6687, we don't serialize query results as JSON
before returning them. This is fine, except for the
query results data source which needs to pass the
data directly to sqlite3, and doesn't know how to
do that with the decimal types that are occasionally
returned by (at least) the PostgreSQL query runner:
https://www.psycopg.org/docs/faq.html#problems-with-type-conversions
* Adds ssl support for prometheus query runner.
- Adds possibilty to upload and use of ssl cert, key and ca file in redash ui
* Extends test cases for prometheus query runner.
- Adds secret attribute to configuration schema.
* Fixes wrong timestamps in different timezones in prometheus' testcases.
- Dynamically calculates timestamps in testcases to be robust in
different timezones.
- Adds now datetime function to make it more testable.
* Fixes timestamp in prometheus' testcases which can be wrong depending on timezone.
---------
Co-authored-by: Masayuki Takahashi <masayuki038@gmail.com>
* Snapshot: 23.11.0-dev
* dataframe_to_result function moved outside python query runner
* added yandex disk query runner
* moved file_extension check
* skip unsupported extensions in schema
* removed unused variable
* added support for xlsx with multiple sheets
* moved pandas-converters to utils file
* added tests
* fixed backend tests
* fixed pandas to redash type conversion
* added more tests
* added tests for pandas
* added tests for pandas converter and yandex disk
* added tests for read_file and multiple sheets
* pandas: do not load if lib is not installed
* added test for yaml read
* fixed test for yaml read
---------
Co-authored-by: github-actions <github-actions@github.com>
Co-authored-by: Guido Petri <18634426+guidopetri@users.noreply.github.com>
* Add column type to a query result of InfluxDB
* Remove an unused import
* Migrate to pytest
* Rename test names to descriptive
---------
Co-authored-by: Guido Petri <18634426+guidopetri@users.noreply.github.com>
* Check all Trino catalogs if no catalog configured
- If catalog is not set, query 'SHOW CATALOGS' to get catalogs
- For each catalog, discover schema, table, columns
- If catalog has a `.`, then quote the catalog name
- New table name returned includes catalog name
- If catalog is set, same result as before
* Use self._handle_run_query_error(error)
* Fix lint problem & format w/ black
* Fix condition so catalog_prefixes are [""] when catalog is set
* Fix for loop / Remove default values for catalog & schema
* Add untested tests for Trino's get_schema
* Black formatting
* Add test_get_schema_catalog_set test & fix tests
* Parameterize the side_effect fn thoroughly
* Get catalogs (not catalog_prefixes) bc confusing
* Trino._get_catalogs added for testability
* Support to find a worksheet by the title for google_spreadsheets
* Add tests for parse_query
* Fix types
* Add tests when finding a worksheet by a title is failed
* Proxy by a wrapper instead of using spreadsheet directly. Do not use gspread in tests
* Add tests. Fix format of quoted strings for titles
* Fix an error format
* Add a newline
* Fix formatting
---------
Co-authored-by: Justin Clift <justin@postgresql.org>
* increase duplicate column names based on the original column name and not the number of duplicates detected
* add fetch columns test for base query runner
---------
Co-authored-by: Omer Lachish <omer@rauchy.net>
Co-authored-by: Guido Petri <18634426+guidopetri@users.noreply.github.com>
* Fix python query runner crashing the GUI due to missing 'rows' and/or 'columns' in the JSON returned data.
* Fix typo of previous commit.
* Throw exception when python query runner has invalid result.
* Update test_python.py
---------
Co-authored-by: YuhengChen <yuheng.chen@imaygou.com>
- A runner supporting the newest versions of ES,
aggregation, nested aggregations and nested fields.
- A runner for the SQL OpenDistro flavor
- A runner for the SQL X-Pack flavor
Co-authored-by: Nicolas Le Manchet <nicolas@lemanchet.fr>
Co-authored-by: wwl717195673 <717195673@qq.com>
ClickHouse query runner splits query into several and execute each query in turn. The result of the last execution is returned. Implementation uses ClickHouse sessions in the HTTP protocol. `session_id` is generated for the first query and then it is used with the subsequent queries (together with the `session_check` parameter).
If query runner gets a success response with empty body from ClickHouse (for example, in case of temporary table creation request) query runner returns empty response.
authored-by: Liubov Ulitina <ulitinalm@vl.ru>
