[
https://issues.apache.org/jira/browse/HIVE-10526?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14519884#comment-14519884
]
Ashutosh Chauhan commented on HIVE-10526:
-----------------------------------------
Is it because CPU & IO are normalized to same domain (query runtime ?) while #
of rows are not. If so, it will be good to add this note as a comment there.
Otherwise, looks good +1.
> CBO (Calcite Return Path): HiveCost epsilon comparison should take row count
> in to account
> ------------------------------------------------------------------------------------------
>
> Key: HIVE-10526
> URL: https://issues.apache.org/jira/browse/HIVE-10526
> Project: Hive
> Issue Type: Sub-task
> Components: CBO
> Reporter: Laljo John Pullokkaran
> Assignee: Laljo John Pullokkaran
> Fix For: 1.2.0
>
> Attachments: HIVE-10526.patch
>
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
