[
https://issues.apache.org/jira/browse/HIVE-15066?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15710427#comment-15710427
]
Prasanth Jayachandran commented on HIVE-15066:
----------------------------------------------
HIVE-7052 have more context to it. Appending values to config object is very
CPU intensive and increases split computation time.
> push down projection is mismatched in HiveInputFormat
> -----------------------------------------------------
>
> Key: HIVE-15066
> URL: https://issues.apache.org/jira/browse/HIVE-15066
> Project: Hive
> Issue Type: Bug
> Components: File Formats
> Affects Versions: 2.1.0
> Reporter: Jeongdae Kim
> Assignee: Jeongdae Kim
> Attachments: HIVE-15066.1.patch
>
>
> if HiveInputFormat.getSplits() is called with more than one table,
> inputFormatClass.getSplits() for the first table will be called with read
> column information for the second table.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
