[
https://issues.apache.org/jira/browse/PIG-3896?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13971759#comment-13971759
]
srinivas commented on PIG-3896:
-------------------------------
Following code snippet will be able duplicate the issue
A= LOAD
B = filter something
Store B
C= filter something
store C
This will create two different pig jobs, in earlier versions it used to create
only on job.
After doing more investigation I noticed that having "STORE" statements should
be in the end of script as opposed to having anywhere in the script in
previous versions.
Is this defect or intentional?
> PigServer - Batch
> ------------------
>
> Key: PIG-3896
> URL: https://issues.apache.org/jira/browse/PIG-3896
> Project: Pig
> Issue Type: Bug
> Components: impl
> Affects Versions: 0.12.0
> Environment: linux
> Reporter: srinivas
> Original Estimate: 0h
> Remaining Estimate: 0h
>
> When using PigServer to execute pig scripts, multiple store statements in a
> single pig script result in multiple pig jobs getting created.
> In earlier versions only on pig job would be created for a pig script even
> there were multiple store statements in the script.
> It looks like batch mode changes affected this
> 0.10.0
> pigServer.setBatchOn();
> pigServer.registerScript(pigScript, params);
> List<ExecJob> pigJobs = pigServer.executeBatch();
> this would create only one pig job
> in version 0.12.0
> pigServer.setBatchOn();
> pigServer.registerScript(pigScript, params);
> List<ExecJob> pigJobs = pigServer.executeBatch(true);
> create multiple pig jobs instead of one and taking longer.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
