You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@hive.apache.org by "László Bodor (Jira)" <ji...@apache.org> on 2026/07/01 09:05:00 UTC

[jira] [Commented] (HIVE-29574) Merge join skew monitor

    [ https://issues.apache.org/jira/browse/HIVE-29574?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18092891#comment-18092891 ] 

László Bodor commented on HIVE-29574:
-------------------------------------

merged to master, thanks a lot [~illiabarbashov225], this was huge! very cool improvement in the skew visibility area! 

> Merge join skew monitor
> -----------------------
>
>                 Key: HIVE-29574
>                 URL: https://issues.apache.org/jira/browse/HIVE-29574
>             Project: Hive
>          Issue Type: Bug
>            Reporter: Illia Barbashov
>            Assignee: Illia Barbashov
>            Priority: Major
>
> When Skew is found in single tez task at merge join, 
>  * Have a threshold (max rows per join key) check
>  * When the join produces more than threshold,
>  ** Either kill the task & while throwing the exception, log the join key responsible for causing skew explode.
>  ** Or at very minimum, log warning about the join key producing huge skew in single task.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)