pabloem commented on a change in pull request #12782: URL: https://github.com/apache/beam/pull/12782#discussion_r493919948
########## File path: sdks/python/apache_beam/runners/dataflow/dataflow_runner_test.py ########## @@ -323,7 +316,9 @@ def test_no_group_by_key_directly_after_bigquery(self): options=PipelineOptions(self.default_properties)) as p: # pylint: disable=expression-not-assigned p | beam.io.Read( - beam.io.BigQuerySource('dataset.faketable')) | beam.GroupByKey() + beam.io.BigQuerySource( + 'dataset.faketable', + use_dataflow_native_source=True)) | beam.GroupByKey() Review comment: for these two tests in dataflow_runner_test, we do not have equivalents for the custom source. This is because these failures do not occur under the custom source: - Custom source is acceptable in streaming because it becomes an SDF that runs fine in streaming - Custom source will not error out on type hints with GBK because the type hints are not set up to fail for this sort of composite transform. I don't know why, but we can file a bug to ensure this is re-enabled. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org