[
https://issues.apache.org/jira/browse/BEAM-2266?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16008184#comment-16008184
]
Thomas Weise commented on BEAM-2266:
------------------------------------
We need to add support for larger bundles, I think that is the only way to get
a better default behavior? Unless we want to modify the example to let the user
influence the shards, like it was done here:
https://github.com/tweise/apex-samples/blob/master/beam-apex-wordcount/src/main/java/com/example/myapexapp/Application.java#L186
> WordCount on ApexRunner writes one word count per file
> ------------------------------------------------------
>
> Key: BEAM-2266
> URL: https://issues.apache.org/jira/browse/BEAM-2266
> Project: Beam
> Issue Type: Bug
> Components: runner-apex
> Reporter: Kenneth Knowles
> Assignee: Thomas Weise
>
> This has been a common issue with most streaming runners. Running on Apex you
> get one "word: <count>" entry per file. Perhaps something smarter could be
> done as a default in such cases. This will also occur with single element
> bundles, though in this case there is a single element per key so single-key
> bundles will result in single-element bundles.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
