[
https://issues.apache.org/jira/browse/BEAM-11715?focusedWorklogId=550462&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-550462
]
ASF GitHub Bot logged work on BEAM-11715:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 09/Feb/21 22:31
Start Date: 09/Feb/21 22:31
Worklog Time Spent: 10m
Work Description: yifanmai commented on pull request #13877:
URL: https://github.com/apache/beam/pull/13877#issuecomment-776288678
> Maybe we could simplify things by only supporting portable job submission?
Even with portable job submission, the subtransform structure is different
(CombineValues is a primitive rather than a composite).
[CombineValuesPTransformOverride](https://github.com/apache/beam/blob/595c25263638b3d5624875e46b69a442cedb09a7/sdks/python/apache_beam/runners/dataflow/ptransform_overrides.py#L128)
performs this substitution in DataflowRunner
[before](https://github.com/apache/beam/blob/595c25263638b3d5624875e46b69a442cedb09a7/sdks/python/apache_beam/runners/dataflow/dataflow_runner.py#L479)
the pipeline is converted to a proto. So the optimizer will still see the
Dataflow-specific subtransform structure.
----------------------------------------------------------------
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:
[email protected]
Issue Time Tracking
-------------------
Worklog Id: (was: 550462)
Time Spent: 10h 20m (was: 10h 10m)
> Combiner packing creates an incorrect proto
> -------------------------------------------
>
> Key: BEAM-11715
> URL: https://issues.apache.org/jira/browse/BEAM-11715
> Project: Beam
> Issue Type: Bug
> Components: sdk-py-core
> Reporter: Chamikara Madhusanka Jayalath
> Assignee: Yifan Mai
> Priority: P1
> Fix For: 2.28.0
>
> Time Spent: 10h 20m
> Remaining Estimate: 0h
>
> Seems like optimization creates a CombinePerKey transform that does not have
> any sub-transforms. We should fix this by preserving the old structure for
> CombinePerKey (GBK+CombineValues composite).
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
