[
https://issues.apache.org/jira/browse/BEAM-8272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16946044#comment-16946044
]
Kenneth Knowles commented on BEAM-8272:
---------------------------------------
State & timers is the way to unify across different bundling choices by the
runner and make window partitioning very simple.
There is often an assumption that in batch mode a runner will create a single
large bundle so you can do this with just startBundle/finishBundle, but that is
not a requirement. It is also not the case that batch implies global window, so
in that implementation strategy you do need to manage keeping disjoint local
state for each window (or fail at construction time in the presence of
non-global windows)
> GroupIntoBatches transform for Go SDK
> -------------------------------------
>
> Key: BEAM-8272
> URL: https://issues.apache.org/jira/browse/BEAM-8272
> Project: Beam
> Issue Type: New Feature
> Components: sdk-go
> Reporter: John Patoch
> Priority: Major
>
> Add a PTransform that batches inputs to a desired batch size. Batches will
> contain only elements of a single key.
> It should offer the same API as its Java counterpart:
> [https://github.com/apache/beam/blob/11a977b8b26eff2274d706541127c19dc93131a2/sdks/java/core/src/main/java/org/apache/beam/sdk/transforms/GroupIntoBatches.java]
>
> And Python counterpart:
> https://github.com/apache/beam/blob/c445fdfdfab4a191aa780210564199f2873f85d8/sdks/python/apache_beam/transforms/util.py#L684
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
