From ec062b6787dbf3a4285cc571014b939e778c0fa3 Mon Sep 17 00:00:00 2001
From: Gabe Lyons <itsgabelyons@gmail.com>
Date: Mon, 7 Feb 2022 14:29:51 -0800
Subject: [PATCH] feat(glue): make ownership configurable in glue source
 (#4078)

---
 metadata-ingestion/source_docs/glue.md        |  1 +
 .../src/datahub/ingestion/source/aws/glue.py  | 20 ++++++++++++-------
 2 files changed, 14 insertions(+), 7 deletions(-)

diff --git a/metadata-ingestion/source_docs/glue.md b/metadata-ingestion/source_docs/glue.md
index a4ddedfb50be8..27f5074f9775f 100644
--- a/metadata-ingestion/source_docs/glue.md
+++ b/metadata-ingestion/source_docs/glue.md
@@ -92,6 +92,7 @@ Note that a `.` is used to denote nested fields in the YAML recipe.
 | `ignore_unsupported_connectors` |          | `True`       | Whether to ignore unsupported connectors. If disabled, an error will be raised.    |
 | `emit_s3_lineage`               |          | `True`       | Whether to emit S3-to-Glue lineage.                                                |
 | `glue_s3_lineage_direction`     |          | `upstream`   | If `upstream`, S3 is upstream to Glue. If `downstream` S3 is downstream to Glue.   |
+| `extract_owners`                |          | `True`       | When enabled, extracts ownership from Glue directly and overwrites existing owners. When disabled, ownership is left empty for datasets.                                                      |
 
 ## Compatibility
 
diff --git a/metadata-ingestion/src/datahub/ingestion/source/aws/glue.py b/metadata-ingestion/src/datahub/ingestion/source/aws/glue.py
index d91a5f2893408..1491992469983 100644
--- a/metadata-ingestion/src/datahub/ingestion/source/aws/glue.py
+++ b/metadata-ingestion/src/datahub/ingestion/source/aws/glue.py
@@ -48,6 +48,7 @@
 
 class GlueSourceConfig(AwsSourceConfig):
 
+    extract_owners: Optional[bool] = True
     extract_transforms: Optional[bool] = True
     underlying_platform: Optional[str] = None
     ignore_unsupported_connectors: Optional[bool] = True
@@ -89,6 +90,7 @@ class GlueSource(Source):
 
     def __init__(self, config: GlueSourceConfig, ctx: PipelineContext):
         super().__init__(ctx)
+        self.extract_owners = config.extract_owners
         self.source_config = config
         self.report = GlueSourceReport()
         self.glue_client = config.glue_client
@@ -612,7 +614,7 @@ def get_workunits(self) -> Iterable[MetadataWorkUnit]:
                     yield dataset_wu
 
     def _extract_record(self, table: Dict, table_name: str) -> MetadataChangeEvent:
-        def get_owner() -> OwnershipClass:
+        def get_owner() -> Optional[OwnershipClass]:
             owner = table.get("Owner")
             if owner:
                 owners = [
@@ -621,11 +623,10 @@ def get_owner() -> OwnershipClass:
                         type=OwnershipTypeClass.DATAOWNER,
                     )
                 ]
-            else:
-                owners = []
-            return OwnershipClass(
-                owners=owners,
-            )
+                return OwnershipClass(
+                    owners=owners,
+                )
+            return None
 
         def get_dataset_properties() -> DatasetPropertiesClass:
             return DatasetPropertiesClass(
@@ -680,7 +681,12 @@ def get_schema_metadata(glue_source: GlueSource) -> SchemaMetadata:
         )
 
         dataset_snapshot.aspects.append(Status(removed=False))
-        dataset_snapshot.aspects.append(get_owner())
+
+        if self.extract_owners:
+            optional_owner_aspect = get_owner()
+            if optional_owner_aspect is not None:
+                dataset_snapshot.aspects.append(optional_owner_aspect)
+
         dataset_snapshot.aspects.append(get_dataset_properties())
         dataset_snapshot.aspects.append(get_schema_metadata(self))