[jira] Commented: (HIVE-291) [Hive] map-side aggregation should be automatically disabled at run-time if it is not turning out to be useful

["Namit Jain (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HIVE-291?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12674893#action_12674893 ] 

Namit Jain commented on HIVE-291:
---------------------------------

tested one big job for correctness

> [Hive] map-side aggregation should be automatically disabled at run-time if it is not turning out to be useful
> --------------------------------------------------------------------------------------------------------------
>
>                 Key: HIVE-291
>                 URL: https://issues.apache.org/jira/browse/HIVE-291
>             Project: Hadoop Hive
>          Issue Type: Bug
>          Components: Query Processor
>            Reporter: Namit Jain
>            Assignee: Namit Jain
>         Attachments: 291.1.txt
>
>
> Map-side aggregation should be automatically disabled at run-time if it is not turning out to be useful.
> If map-side aggregation is not reducing the number of output rows, it is a drain on the mapper, since it is consuming memory and performing unnecessary hash lookups

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.