[
https://issues.apache.org/jira/browse/PIG-4702?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14959391#comment-14959391
]
Rohini Palaniswamy commented on PIG-4702:
-----------------------------------------
bq. Why Native_2 failing by PIG-4574?
If just load is followed by order by, partitioning and sampling are done by
reading data again. For that plan is cloned. Usually, POLoad.getInputs() is
null. But in case of native jobs it has the native job plan and it throws
errors while cloning.
> Load once for sampling and partitioning in order by for certain LoadFuncs
> -------------------------------------------------------------------------
>
> Key: PIG-4702
> URL: https://issues.apache.org/jira/browse/PIG-4702
> Project: Pig
> Issue Type: Improvement
> Components: tez
> Reporter: Rohini Palaniswamy
> Assignee: Rohini Palaniswamy
> Fix For: 0.16.0
>
> Attachments: PIG-4702-1.patch
>
>
> For HBase and Accumulo, it will be more efficient on IO to have the data
> written to disk instead of reading from them again.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
