[
https://issues.apache.org/jira/browse/FLINK-31821?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17712991#comment-17712991
]
Martijn Visser commented on FLINK-31821:
----------------------------------------
[~lenault] Given that there are no replies yet on
https://lists.apache.org/thread/y6y8f39dxg644c5hvl2l86q8jv2z0rmf, it appears
that a) users need more time or b) that there isn't so much interest in the
feature. Adding a feature like this touches on a lot of public interfaces, so
it needs a broad discussion and consensus.
> FlinkSQL set parallelism for each operator.
> -------------------------------------------
>
> Key: FLINK-31821
> URL: https://issues.apache.org/jira/browse/FLINK-31821
> Project: Flink
> Issue Type: New Feature
> Components: API / Core
> Affects Versions: 1.16.1
> Reporter: hanqing
> Priority: Critical
> Attachments: image-2023-04-17-17-31-34-709.png
>
> Original Estimate: 168h
> Remaining Estimate: 168h
>
> Currently, FlinkSQL can set a unified parallelism in the job,it cannot set
> parallelism for each operator.
> This can cause resource waste On the occasion of high parallelism and small
> data volume.there may also be too many small file for writing HDFS Scene.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
