Mackenzie created BEAM-5466:
-------------------------------
Summary: Cannot deploy job to Dataflow with RuntimeValueProvider
Key: BEAM-5466
URL: https://issues.apache.org/jira/browse/BEAM-5466
Project: Beam
Issue Type: Bug
Components: runner-dataflow, sdk-py-core
Affects Versions: 2.6.0
Environment: Python 2.7
Reporter: Mackenzie
Assignee: Henning Rohde
I cannot deploy an apache beam job to Cloud Dataflow that contains runtime
value parameters.
The standard use case is with Cloud Dataflow Templates which use
RuntimeValueProvider to get template parameters.
When trying to call `get` on the parameter, I always get an error like:
{noformat}
apache_beam.error.RuntimeValueProviderError: RuntimeValueProvider(option:
myparam, type: str, default_value: 'defalut-value').get() not called from a
runtime context
{noformat}
A minimal example:
{code:java}
class UserOptions(PipelineOptions):
@classmethod
def _add_argparse_args(cls, parser):
parser.add_value_provider_argument('--myparam', type=str,
default='default-value')
def run(argv=None):
parser = argparse.ArgumentParser()
known_args, pipeline_args = parser.parse_known_args(argv)
pipeline_options = PipelineOptions(pipeline_args)
pipeline_options.view_as(SetupOptions).save_main_session = True
google_cloud_options = pipeline_options.view_as(GoogleCloudOptions)
# insert google cloud options here, or pass them in arguments
standard_options = pipeline_options.view_as(StandardOptions)
standard_options.runner = 'DataflowRunner'
user_options = pipeline_options.view_as(UserOptions)
p = beam.Pipeline(options=pipeline_options)
param = user_options.myparam.get() # This line is the issue
result = p.run()
result.wait_until_finish()
if __name__ == '__main__':
run()
{code}
I would expect that the runtime context would be ignored when running the
script locally.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
