[
https://issues.apache.org/jira/browse/HADOOP-4565?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12664704#action_12664704
]
Joydeep Sen Sarma commented on HADOOP-4565:
-------------------------------------------
sorry for bringing this up so late - but should we be concerned that we are
going to only use rack locality?
the reason i am bringing this up is that from a usage point of view - this may
complicate how we use this. ie. - ideally we would just always use the CMFIF
with some target split sizes - but now there is a question in my head on
whether we should use CMFIF only in more extreme cases (like too many small
files or too much data and high selectivity) - because otherwise we might use
node locality for not much gain (since in majority of cases - the current
inputformats are just fine)
thoughts?
(still making my way through the code one last time).
> MultiFileInputSplit can use data locality information to create splits
> ----------------------------------------------------------------------
>
> Key: HADOOP-4565
> URL: https://issues.apache.org/jira/browse/HADOOP-4565
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Reporter: dhruba borthakur
> Assignee: dhruba borthakur
> Attachments: CombineMultiFile.patch, CombineMultiFile2.patch,
> CombineMultiFile3.patch, CombineMultiFile4.patch, CombineMultiFile5.patch,
> CombineMultiFile7.patch
>
>
> The MultiFileInputFormat takes a set of paths and creates splits based on
> file sizes. Each splits contains a few files an each split are roughly equal
> in size. It would be efficient if we can extend this InputFormat to create
> splits such each all the blocks in one split and either node-local or
> rack-local.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
