[
https://issues.apache.org/jira/browse/HIVE-20048?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16529407#comment-16529407
]
Gopal V commented on HIVE-20048:
--------------------------------
-1 on this, because local tasks aren't vectorized (& HS2 has to do more work,
which limits concurrency).
This would actually slow down queries in everything other than MRv2 (where we
spin up a 2nd job to do the aggr task, which is what this bypasses).
> Consider enabling hive.fetch.task.aggr by default
> -------------------------------------------------
>
> Key: HIVE-20048
> URL: https://issues.apache.org/jira/browse/HIVE-20048
> Project: Hive
> Issue Type: Task
> Components: Physical Optimizer
> Reporter: Ashutosh Chauhan
> Priority: Major
>
> This optimization is in code base for long time. We shall consider enabling
> it by default.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
