fix(ingest/snowflake): Allow SnowflakeObjectAccessEntry.objectId to be None (#7601)

Co-authored-by: Pedro Silva <[email protected]>

Author: asikowitz

metadata-ingestion/src/datahub/ingestion/source/snowflake/snowflake_usage_v2.py

@@ -66,8 +66,9 @@ class SnowflakeColumnReference(PermissiveModel):
 class SnowflakeObjectAccessEntry(PermissiveModel):
     columns: Optional[List[SnowflakeColumnReference]]
     objectDomain: str
-    objectId: int
     objectName: str
+    # Seems like it should never be null, but in practice have seen null objectIds
+    objectId: Optional[int]
     stageKind: Optional[str]
 
 

metadata-ingestion/tests/unit/test_snowflake_source.py

@@ -11,6 +11,9 @@
     SnowflakeCloudProvider,
 )
 from datahub.ingestion.source.snowflake.snowflake_config import SnowflakeV2Config
+from datahub.ingestion.source.snowflake.snowflake_usage_v2 import (
+    SnowflakeObjectAccessEntry,
+)
 from datahub.ingestion.source.snowflake.snowflake_v2 import SnowflakeV2Source
 
 
@@ -546,3 +549,16 @@ def test_unknown_cloud_region_from_snowflake_region_id():
             "somecloud_someregion"
         )
     assert "Unknown snowflake region" in str(e)
+
+
+def test_snowflake_object_access_entry_missing_object_id():
+    SnowflakeObjectAccessEntry(
+        **{
+            "columns": [
+                {"columnName": "A"},
+                {"columnName": "B"},
+            ],
+            "objectDomain": "View",
+            "objectName": "SOME.OBJECT.NAME",
+        }
+    )