[
https://issues.apache.org/jira/browse/BEAM-10010?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17112672#comment-17112672
]
Brian Hulette commented on BEAM-10010:
--------------------------------------
test_tagged_join reliably produces the following error from Python at execution
time:
{code}
apache_beam/runners/portability/fn_api_runner/fn_runner.py:859: in
process_bundle
output.transform_id).append(output.data)
apache_beam/runners/portability/fn_api_runner/execution.py:591: in get_buffer
coder_impl=self.get_input_coder_impl(transform_id))
apache_beam/runners/portability/fn_api_runner/execution.py:555: in
get_input_coder_impl
return self.get_coder_impl(coder_id)
apache_beam/runners/portability/fn_api_runner/execution.py:574: in
get_coder_impl
return self.execution_context.pipeline_context.coders[coder_id].get_impl()
apache_beam/runners/pipeline_context.py:137: in __getitem__
return self.get_by_id(id)
apache_beam/runners/pipeline_context.py:103: in get_by_id
self._id_to_proto[id], self._pipeline_context)
apache_beam/coders/coders.py:373: in from_runner_api
for c in coder_proto.component_coder_ids
apache_beam/coders/coders.py:373: in <listcomp>
for c in coder_proto.component_coder_ids
apache_beam/runners/pipeline_context.py:103: in get_by_id
self._id_to_proto[id], self._pipeline_context)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
cls = <class 'apache_beam.coders.coders.Coder'>
coder_proto = spec {
urn: "beam:coders:javasdk:0.1"
payload:
"\202SNAPPY\000\000\000\000\001\000\000\000\001\000\000\0060\241\02...000\030\235%\r8\005\362\030;xpvr\000\036>\355\000\301\201-\313(Row,\366\306j\2507,G\005X4\006schemaq\000~\000\002xp"
}
context = <apache_beam.runners.pipeline_context.PipelineContext object at
0x7fe6b3e4bfd0>
@classmethod
def from_runner_api(cls, coder_proto, context):
# type: (Type[CoderT], beam_runner_api_pb2.Coder, PipelineContext) ->
CoderT
"""Converts from an FunctionSpec to a Fn object.
Prefer registering a urn with its parameter type and constructor.
"""
if (context.allow_proto_holders and
coder_proto.spec.urn not in cls._known_urns):
# We hold this in proto form since there's no coder available in Python
# SDK.
# This is potentially a coder that is only available in an external SDK.
return ExternalCoder(coder_proto)
else:
> parameter_type, constructor = cls._known_urns[coder_proto.spec.urn]
E KeyError: 'beam:coders:javasdk:0.1'
apache_beam/coders/coders.py:369: KeyError
{code}
> Test Python SqlTransform on fn_api_runner
> -----------------------------------------
>
> Key: BEAM-10010
> URL: https://issues.apache.org/jira/browse/BEAM-10010
> Project: Beam
> Issue Type: Improvement
> Components: sdk-py-core
> Reporter: Brian Hulette
> Assignee: Brian Hulette
> Priority: P2
>
> It should be possible to run with the fn_api_runner.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
