refreshing snowflake schema w/o waking cluster (#4285)

* refreshing snowflake schema w/o waking cluster

Have also added a new internal method to not select a
warehouse while executing query
Using 'show columns' to fetch database schema instead of
executing a select query in information schema
show columns does not require a warehouse to run

* modularising snowflake code to avoid repetitions

fixing internal function syntax and avoiding
code repetition

* removing user object in snowflake schema query

redash/query_runner/snowflake.py

@@ -67,7 +67,7 @@ class Snowflake(BaseQueryRunner):
             return TYPE_FLOAT
         return t
 
-    def run_query(self, query, user):
+    def _get_connection(self):
         region = self.configuration.get('region')
 
         # for us-west we don't need to pass a region (and if we do, it fails to connect)
@@ -81,6 +81,19 @@ class Snowflake(BaseQueryRunner):
             region=region
         )
 
+        return connection
+
+    def _parse_results(self, cursor):
+        columns = self.fetch_columns(
+            [(i[0], self.determine_type(i[1], i[5])) for i in cursor.description])
+        rows = [dict(zip((column['name'] for column in columns), row))
+                for row in cursor]
+
+        data = {'columns': columns, 'rows': rows}
+        return data
+
+    def run_query(self, query, user):
+        connection = self._get_connection()
         cursor = connection.cursor()
 
         try:
@@ -90,12 +103,7 @@ class Snowflake(BaseQueryRunner):
 
             cursor.execute(query)
 
-            columns = self.fetch_columns(
-                    [(i[0], self.determine_type(i[1], i[5])) for i in cursor.description])
-            rows = [dict(zip((column['name'] for column in columns), row))
-                    for row in cursor]
-
-            data = {'columns': columns, 'rows': rows}
+            data = self._parse_results(cursor)
             error = None
             json_data = json_dumps(data)
         finally:
@@ -104,30 +112,42 @@ class Snowflake(BaseQueryRunner):
 
         return json_data, error
 
+    def _run_query_without_warehouse(self, query):
+        connection = self._get_connection()
+        cursor = connection.cursor()
+
+        try:
+            cursor.execute("USE {}".format(self.configuration['database']))
+
+            cursor.execute(query)
+
+            data = self._parse_results(cursor)
+            error = None
+        finally:
+            cursor.close()
+            connection.close()
+
+        return data, error
+
     def get_schema(self, get_stats=False):
         query = """
-        SELECT col.table_schema,
-               col.table_name,
-               col.column_name
-        FROM {database}.information_schema.columns col
-        WHERE col.table_schema <> 'INFORMATION_SCHEMA'
+        SHOW COLUMNS IN DATABASE {database}
         """.format(database=self.configuration['database'])
 
-        results, error = self.run_query(query, None)
+        results, error = self._run_query_without_warehouse(query)
 
         if error is not None:
             raise Exception("Failed getting schema.")
 
         schema = {}
-        results = json_loads(results)
-
         for row in results['rows']:
-            table_name = '{}.{}'.format(row['TABLE_SCHEMA'], row['TABLE_NAME'])
+            if row['kind'] == 'COLUMN':
+                table_name = '{}.{}'.format(row['schema_name'], row['table_name'])
 
-            if table_name not in schema:
-                schema[table_name] = {'name': table_name, 'columns': []}
+                if table_name not in schema:
+                    schema[table_name] = {'name': table_name, 'columns': []}
 
-            schema[table_name]['columns'].append(row['COLUMN_NAME'])
+                schema[table_name]['columns'].append(row['column_name'])
 
         return list(schema.values())