[
https://issues.apache.org/jira/browse/HIVE-2339?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13576683#comment-13576683
]
Yin Huai commented on HIVE-2339:
--------------------------------
should we reopen this one? From the discussion of HIVE-2340, seems HIVE-1989
has not completely resolve the issue. Also, HIVE-2206 needs to use mapping of
column names to track common keys among ReduceSinkOperators. Since the mapping
of column names of keys are missing, if "hive.map.aggr=false", the current
patch of HIVE-2206 cannot detect common keys. "hive.map.aggr=true" will not be
a problem since I still generate a reduce-side aggregation which is not in the
plan tree and thus, will not go through CP optimization
> Preserve RS key columns in columnExprMap after CP optimization, which might
> be useful to other optimizers
> ---------------------------------------------------------------------------------------------------------
>
> Key: HIVE-2339
> URL: https://issues.apache.org/jira/browse/HIVE-2339
> Project: Hive
> Issue Type: Sub-task
> Components: Query Processor
> Reporter: Navis
> Assignee: Navis
> Priority: Trivial
> Fix For: 0.8.0
>
> Attachments: HIVE-2339.1.patch
>
>
> In ColumnPrunerProcFactory#pruneReduceSinkOperator, only VALUE parts are
> retained from columnExprMap. Doesn't anyone want KEY parts to retained,
> either? In my case, it was very useful for backtracking column names and
> removing RS in *-RS-*-RS-GBY case.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
