Merge branch 'main' into 1.10.latest

Author: ericj-db

CHANGELOG.md

@@ -1,4 +1,17 @@
-## dbt-databricks 1.10.8 (TBD)
+## dbt-databricks 1.10.9 (TBD)
+
+## dbt-databricks 1.10.8 (August 4, 2025)
+
+### Features
+- Support insert_overwrite incremental strategy for SQL warehouses ([1025](https://github.com/databricks/dbt-databricks/issues/1025))
+
+### Fixes
+- Add fallback logic for known error types for `DESCRIBE TABLE EXTENDED .. AS JSON` for better reliability ([1128](https://github.com/databricks/dbt-databricks/issues/1128))
+- Fix no-op logic for views that is causing some incremental materializations to be skipped ([1122](https://github.com/databricks/dbt-databricks/issues/1122))
+- Fix check constraints keep getting replaced [issue-1109](https://github.com/databricks/dbt-databricks/issues/1109)
+
+### Under the Hood
+- Simplify connection management to align with base adapter. Connections are no longer cached per-thread
 
 ## dbt-databricks 1.10.7 (July 31, 2025)
 

dbt/adapters/databricks/__version__.py

@@ -1 +1 @@
-version = "1.10.7"
+version = "1.10.8"

dbt/adapters/databricks/behaviors/columns.py

@@ -1,5 +1,6 @@
 from abc import ABC, abstractmethod
 
+from dbt_common.exceptions import DbtDatabaseError
 from dbt_common.utils.dict import AttrDict
 
 from dbt.adapters.databricks.column import DatabricksColumn
@@ -37,13 +38,20 @@ def get_columns_in_relation(
             rows = cls._get_columns_with_comments(adapter, relation, "get_columns_comments")
             return cls._parse_columns(rows)
         else:
-            result = cls._get_columns_with_comments(
-                adapter, relation, "get_columns_comments_as_json"
-            )
-            if not result:
-                return []
-            json_metadata = result[0]["json_metadata"]
-            return DatabricksColumn.from_json_metadata(json_metadata)
+            try:
+                result = cls._get_columns_with_comments(
+                    adapter, relation, "get_columns_comments_as_json"
+                )
+                if not result:
+                    return []
+                json_metadata = result[0]["json_metadata"]
+                return DatabricksColumn.from_json_metadata(json_metadata)
+            except DbtDatabaseError as ex:
+                # Fall back to legacy logic if the error is due to AS JSON not being supported
+                # for the current runtime or relation type (e.g. foreign table)
+                if "PARSE_SYNTAX_ERROR" in ex.msg or "UNSUPPORTED_FEATURE" in ex.msg:
+                    return cls.get_columns_in_relation(adapter, relation, True)
+                raise ex
 
     @classmethod
     def _parse_columns(cls, rows: list[AttrDict]) -> list[DatabricksColumn]: