[ 
https://issues.apache.org/jira/browse/FLINK-20611?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Suhan Mao updated FLINK-20611:
------------------------------
    Description: 
Currently, mini batch does not support window aggregate. In our company, many 
jobs run with window aggregate and suffer from large amounts of intermediate 
shuffle. If we can enable mini batch on window aggregate, the performance will 
be improved.

This issue aims to support mini batch in window aggregate operator both in SQL 
and Table API. Simple tumble window will be supported in the first place.

 

We still use below configurations, but they take effect only inside each single 
window.

{{table.exec.mini-batch.enabled}}

{{table.exec.mini-batch.allow-latency}}

{{table.exec.mini-batch.size}}

 

  was:
Currently, mini batch does not support window aggregate. In our company, many 
jobs run with window aggregate and suffer from large amounts of intermediate 
shuffle. If we can enable mini batch on window aggregate, the performance will 
be improved.

This issue aims to support mini batch in window aggregate operator both in SQL 
and Table API. Simple tumble window will be supported in the first place.

 

We still use below configurations, but they take effect only inside each single 
window.

{{```}}

{{table.exec.mini-batch.enabled}}

{{table.exec.mini-batch.allow-latency}}

{{table.exec.mini-batch.size}}

{{```}}

 


> Support mini batch for window aggregate operation
> -------------------------------------------------
>
>                 Key: FLINK-20611
>                 URL: https://issues.apache.org/jira/browse/FLINK-20611
>             Project: Flink
>          Issue Type: New Feature
>          Components: Table SQL / API
>            Reporter: Suhan Mao
>            Priority: Major
>
> Currently, mini batch does not support window aggregate. In our company, many 
> jobs run with window aggregate and suffer from large amounts of intermediate 
> shuffle. If we can enable mini batch on window aggregate, the performance 
> will be improved.
> This issue aims to support mini batch in window aggregate operator both in 
> SQL and Table API. Simple tumble window will be supported in the first place.
>  
> We still use below configurations, but they take effect only inside each 
> single window.
> {{table.exec.mini-batch.enabled}}
> {{table.exec.mini-batch.allow-latency}}
> {{table.exec.mini-batch.size}}
>  



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

Reply via email to