[
https://issues.apache.org/jira/browse/HIVE-9966?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14361467#comment-14361467
]
Ashutosh Chauhan commented on HIVE-9966:
----------------------------------------
>From MapJoinOp traverse on big table side. If TS is encountered, then custom
>edge is needed, else not.
> Get rid of customBucketMapJoin field from MapJoinDesc
> -----------------------------------------------------
>
> Key: HIVE-9966
> URL: https://issues.apache.org/jira/browse/HIVE-9966
> Project: Hive
> Issue Type: Task
> Components: Query Planning, Tez
> Reporter: Ashutosh Chauhan
> Assignee: Ashutosh Chauhan
>
> Currently, its used to determine whether BMJ is running in mapper or reducer
> in ReduceSinkMapJoinProc rule. But this determination can be made locally by
> examining operator tree in the rule.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
