[BEAM-10172] BigQuerySource external data source support in non-US regions

• Type: Bug
• Status: Open
• Priority: P3
• Resolution: Unresolved
• Affects Version/s: 2.18.0
• Component/s: io-py-gcp
• Environment: DirectRunner, DataflowRunner
• Assignee: Unassigned
• Reporter: Boris Shilov
• Created: 02/Jun/20 11:54
• Updated: 04/Jun/22 16:22

Description

I am attempting to query an external data source, a MySQL database that is exposed via the BigQuery API, located in the EU region. I have the following format query string:

query = """
    SELECT * 
    FROM EXTERNAL_QUERY("my-project-one-253518.eu.external-source", 
    "SELECT * FROM my schema.mytable;");
    """

And the following pipeline instantiation:

    pcoll = p | "Load " + name >> beam.io.Read(
        beam.io.BigQuerySource(query=query, use_standard_sql=True)
    )

When run this, I see the following output:

WARNING:root:Dataset my-project-two:temp_dataset_f07dd1398b0443edaa67c360f5be6958 does not exist so we will create it as temporary with location=None
ERROR:root:Exception at bundle <apache_beam.runners.direct.bundle_factory._Bundle object at 0x127124640>, due to an exception.
 Traceback (most recent call last):
  File "venv/lib/python3.7/site-packages/apache_beam/runners/direct/executor.py", line 345, in call
    finish_state)
  File "venv/lib/python3.7/site-packages/apache_beam/runners/direct/executor.py", line 385, in attempt_call
    result = evaluator.finish_bundle()
  File "venv/lib/python3.7/site-packages/apache_beam/runners/direct/transform_evaluator.py", line 323, in finish_bundle
    bundles = _read_values_to_bundles(reader)
  File "venv/lib/python3.7/site-packages/apache_beam/runners/direct/transform_evaluator.py", line 310, in _read_values_to_bundles
    read_result = [GlobalWindows.windowed_value(e) for e in reader]
  File "venv/lib/python3.7/site-packages/apache_beam/runners/direct/transform_evaluator.py", line 310, in <listcomp>
    read_result = [GlobalWindows.windowed_value(e) for e in reader]
  File "venv/lib/python3.7/site-packages/apache_beam/io/gcp/bigquery_tools.py", line 937, in __iter__
    flatten_results=self.flatten_results):
  File "venv/lib/python3.7/site-packages/apache_beam/io/gcp/bigquery_tools.py", line 710, in run_query
    page_token, location=location)
  File "venv/lib/python3.7/site-packages/apache_beam/utils/retry.py", line 209, in wrapper
    return fun(*args, **kwargs)
  File "venv/lib/python3.7/site-packages/apache_beam/io/gcp/bigquery_tools.py", line 384, in _get_query_results
    response = self.client.jobs.GetQueryResults(request)
  File "venv/lib/python3.7/site-packages/apache_beam/io/gcp/internal/clients/bigquery/bigquery_v2_client.py", line 312, in GetQueryResults
    config, request, global_params=global_params)
  File "venv/lib/python3.7/site-packages/apitools/base/py/base_api.py", line 731, in _RunMethod
    return self.ProcessHttpResponse(method_config, http_response, request)
  File "venv/lib/python3.7/site-packages/apitools/base/py/base_api.py", line 737, in ProcessHttpResponse
    self.__ProcessHttpResponse(method_config, http_response, request))
  File "venv/lib/python3.7/site-packages/apitools/base/py/base_api.py", line 604, in __ProcessHttpResponse
    http_response, method_config=method_config, request=request)
apitools.base.py.exceptions.HttpBadRequestError: HttpError accessing <https://www.googleapis.com/bigquery/v2/projects/my-project-two/queries/636272a8e026434d85200b3f14f719ed?alt=json&location=US&maxResults=10000>: response: <{'vary': 'Origin, X-Origin, Referer', 'content-type': 'application/json; charset=UTF-8', 'date': 'Tue, 02 Jun 2020 11:29:27 GMT', 'server': 'ESF', 'cache-control': 'private', 'x-xss-protection': '0', 'x-frame-options': 'SAMEORIGIN', 'x-content-type-options': 'nosniff', 'alt-svc': 'h3-27=":443"; ma=2592000,h3-25=":443"; ma=2592000,h3-T050=":443"; ma=2592000,h3-Q050=":443"; ma=2592000,h3-Q049=":443"; ma=2592000,h3-Q048=":443"; ma=2592000,h3-Q046=":443"; ma=2592000,h3-Q043=":443"; ma=2592000,quic=":443"; ma=2592000; v="46,43"', 'transfer-encoding': 'chunked', 'status': '400', 'content-length': '354', '-content-encoding': 'gzip'}>, content <{
  "error": {
    "code": 400,
    "message": "Cannot read and write in different locations: source: EU, destination: US",
    "errors": [
      {
        "message": "Cannot read and write in different locations: source: EU, destination: US",
        "domain": "global",
        "reason": "invalid"
      }
    ],
    "status": "INVALID_ARGUMENT"
  }
}

Which likely indicates to me that the logic Beam uses to impute the zone in which to create the temporary dataset fails when confronted with the special syntax for external queries. Therefore it seems like the zone should be exposed as a parameter of BigQuerySource.