You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@pig.apache.org by "Daniel Dai (JIRA)" <ji...@apache.org> on 2013/01/21 23:46:12 UTC

[jira] [Commented] (PIG-3124) Push FLATTENs After FILTERs If Possible

    [ https://issues.apache.org/jira/browse/PIG-3124?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13559182#comment-13559182 ] 

Daniel Dai commented on PIG-3124:
---------------------------------

Seems it has been optimized by rule FilterAboveForeach. The test case testForeachFilter is confusing, we should put the following remark: PushDownForEachFlatten does not apply here because Pig decide to leave it to rule FilterAboveForeach.
                
> Push FLATTENs After FILTERs If Possible
> ---------------------------------------
>
>                 Key: PIG-3124
>                 URL: https://issues.apache.org/jira/browse/PIG-3124
>             Project: Pig
>          Issue Type: Improvement
>          Components: impl
>    Affects Versions: 0.10.0
>            Reporter: Nick White
>            Assignee: Nick White
>             Fix For: 0.12
>
>         Attachments: PIG-3124.0.patch
>
>
> When optimizing a logical plan, it's safe to push a FLATTEN after a FILTER if the columns being flattened don't occur in the expression that the filter is being done on. When the FILTER comes first the FLATTEN generates fewer rows (usually), and so is more efficient.

--
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