[jira] [Commented] (FLINK-19322) Specified parallelism in create table operation

["Benchao Li (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/FLINK-19322?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17199294#comment-17199294 ] 

Benchao Li commented on FLINK-19322:
------------------------------------

This is currently discussed in [http://apache-flink-mailing-list-archive.1008284.n3.nabble.com/DISCUSS-Introduce-SupportsParallelismReport-and-SupportsStatisticsReport-for-Hive-and-Filesystem-td43531.html]

And also there is another duplicated discussion: [http://apache-flink-mailing-list-archive.1008284.n3.nabble.com/DISCUSS-Support-source-sink-parallelism-config-in-Flink-sql-tc44556.html]

> Specified parallelism in create table operation
> -----------------------------------------------
>
>                 Key: FLINK-19322
>                 URL: https://issues.apache.org/jira/browse/FLINK-19322
>             Project: Flink
>          Issue Type: Improvement
>          Components: Connectors / Common, Table SQL / API
>    Affects Versions: 1.12.0
>            Reporter: YufeiLiu
>            Priority: Major
>
> Flink SQL job only has global parallelism, and all operators share the same parallelism. But in many cases like using kafka connector, I want to set table parallelism separately to control resource usage.
> * Source & Sink topic has different partition size.
> * Consume multi topic with different throughput.
> * Increase sink parallelism to insure write speed.
> I think it would be better if we can specified parallelism property in create table operation, and the actual implementation depends on different connectors.
> The create table operation may like this:
> {code:sql}
> CREATE TABLE source (
>   ...
> ) WITH (
>   'connector' = 'kafka',
>   'parallelism' = '2'
> )
> {code}



--
This message was sent by Atlassian Jira
(v8.3.4#803005)