[
https://issues.apache.org/jira/browse/BEAM-1611?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15893244#comment-15893244
]
ASF GitHub Bot commented on BEAM-1611:
--------------------------------------
GitHub user kennknowles opened a pull request:
https://github.com/apache/beam/pull/2147
[BEAM-1611] Explicitly GBK before stateful ParDo in Dataflow batch
Be sure to do all of the following to help us incorporate your contribution
quickly and easily:
- [ ] Make sure the PR title is formatted like:
`[BEAM-<Jira issue #>] Description of pull request`
- [ ] Make sure tests pass via `mvn clean verify`. (Even better, enable
Travis-CI on your fork and ensure the whole test matrix passes).
- [ ] Replace `<Jira issue #>` in the title with the actual Jira issue
number, if there is one.
- [ ] If this contribution is large, please file an Apache
[Individual Contributor License
Agreement](https://www.apache.org/licenses/icla.txt).
---
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/kennknowles/beam state-issue
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/beam/pull/2147.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #2147
----
commit 2cfcac27bbf961f3dd017701b0eee9160a4a847d
Author: Kenneth Knowles <[email protected]>
Date: 2017-03-02T22:29:56Z
In Dataflow batch, explicitly GroupByKey before a stateful ParDo
commit 15a781cbb280341e863bccac04332965797572a1
Author: Kenneth Knowles <[email protected]>
Date: 2017-03-02T22:48:26Z
Add test for deduping by key with stateful ParDo
commit be4271c3a5cadbb92dfbb707fb1f50a486bca827
Author: Kenneth Knowles <[email protected]>
Date: 2017-03-02T22:57:42Z
HACK: use ParDo multi until it is the only thing
----
> Stateful ParDo broken in Dataflow batch
> ---------------------------------------
>
> Key: BEAM-1611
> URL: https://issues.apache.org/jira/browse/BEAM-1611
> Project: Beam
> Issue Type: Bug
> Components: runner-dataflow
> Reporter: Kenneth Knowles
> Assignee: Kenneth Knowles
> Priority: Critical
> Fix For: 0.6.0
>
>
> The needed shuffle is not performed in batch mode. Confirmed that it is
> performed in streaming.
> http://stackoverflow.com/questions/42564604/stateful-pardo-not-working-on-dataflow-runner/42566967#42566967
> Filing here as the quickest fix is to work around with transform overrides in
> Beam, analogous to those in the direct runner.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
