[
https://issues.apache.org/jira/browse/TEZ-624?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13877031#comment-13877031
]
Bikas Saha commented on TEZ-624:
--------------------------------
Changes vertexindex javadoc. Opened TEZ-749 to maintain vertex order based on
user addition into the DAG.
Unique identifies change is not that obvious. In this case, the MROutput can
get away with it because of the task id based hierarchy being created. In
general, TezTaskContext.getUniqueIdentifier() will be different from
TezTaskOutputContext.getUniqueIdentifier() if we want to based it on the
vertex,task,attempt,output indices. If needed we can change the current impl.
Added InputIndex.
We can change to duplicating the code if perf becomes and issue which likely
wont since the main configuration object has already been created. Until its an
issue I'd rather stay behind the API instead of duplicating the lookup and
trimming code. In a 2 stage MR job this should probably be fine.
Thanks for the review. Committing last patch.
> Fix MROutput to support multiple outputs to the same output location
> --------------------------------------------------------------------
>
> Key: TEZ-624
> URL: https://issues.apache.org/jira/browse/TEZ-624
> Project: Apache Tez
> Issue Type: Bug
> Reporter: Hitesh Shah
> Assignee: Bikas Saha
> Attachments: TEZ-624.1.patch, TEZ-624.2.patch, TEZ-624.3.patch,
> TEZ-624.4.patch, TEZ-624.5.patch
>
>
> For users using multiple MROutput instances on a single vertex
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
