[
https://issues.apache.org/jira/browse/FLINK-1622?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14355153#comment-14355153
]
ASF GitHub Bot commented on FLINK-1622:
---------------------------------------
Github user fhueske commented on the pull request:
https://github.com/apache/flink/pull/466#issuecomment-78091524
I have to correct myself. A combiner should of course be called on groups
of records. Therefore, calling it on a Grouping makes absolute sense. However,
the semantics of the Grouping are different for Combine and Reduce. For Reduce
a full Grouping is established with repartitioning and full sort, whereas for
Combine only a partial local grouping should be created.
> Add GroupReducePartial Operator
> -------------------------------
>
> Key: FLINK-1622
> URL: https://issues.apache.org/jira/browse/FLINK-1622
> Project: Flink
> Issue Type: Sub-task
> Affects Versions: 0.9
> Reporter: Aljoscha Krettek
>
> This does what a Reduce or GroupReduce Operator does, except it is only
> performed on a local partition.
> This is also similar to an explicit combine that can output a type that is
> different from the input.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
