acryl-datahub 0.15.0rc22__py3-none-any.whl → 0.15.0rc24__py3-none-any.whl

datahub/ingestion/source/pulsar.py

@@ -89,7 +89,16 @@ class PulsarSchema:
             logger.error(f"Invalid JSON schema: {schema_data}. Error: {str(e)}")
             avro_schema = {}
 
-        self.schema_name = avro_schema.get("namespace") + "." + avro_schema.get("name")
+        self.schema_name = "null"
+        if avro_schema.get("namespace") and avro_schema.get("name"):
+            self.schema_name = (
+                avro_schema.get("namespace") + "." + avro_schema.get("name")
+            )
+        elif avro_schema.get("namespace"):
+            self.schema_name = avro_schema.get("namespace")
+        elif avro_schema.get("name"):
+            self.schema_name = avro_schema.get("name")
+
         self.schema_description = avro_schema.get("doc")
         self.schema_type = schema.get("type")
         self.schema_str = schema.get("data")

datahub/ingestion/source/snowflake/snowflake_lineage_v2.py

@@ -414,9 +414,13 @@ class SnowflakeLineageExtractor(SnowflakeCommonMixin, Closeable):
         except Exception as e:
             self.report.num_upstream_lineage_edge_parsing_failed += 1
             upstream_tables = db_row.get("UPSTREAM_TABLES")
+            downstream_table = db_row.get("DOWNSTREAM_TABLE_NAME")
             self.structured_reporter.warning(
                 "Failed to parse lineage edge",
-                context=f"Upstreams: {upstream_tables} Downstreams: {db_row.get('DOWNSTREAM_TABLE_NAME')}",
+                # Tricky: sometimes the full row data is too large, and so the context
+                # message gets truncated. By pulling out the upstreams and downstream
+                # list, we can at least get the important fields if truncation does occur.
+                context=f"Upstreams: {upstream_tables} Downstream: {downstream_table} Full row: {db_row}",
                 exc=e,
             )
             return None