[
https://issues.apache.org/jira/browse/HAMA-774?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13701922#comment-13701922
]
Martin Illecker commented on HAMA-774:
--------------------------------------
Thanks, currently I'm using the same functionality like it were a directory:
{code}
if ((job.get("bsp.input.dir") != null) || (job.get("bsp.join.expr") != null)) {
// Create the splits for the job
{code}
It works, but I get the following output for only two input files?
{code}
INFO bsp.FileInputFormat: Total input paths to process : 1
INFO bsp.FileInputFormat: Total input paths to process : 1
INFO bsp.FileInputFormat: Total input paths to process : 1
INFO bsp.FileInputFormat: Total input paths to process : 1
{code}
Also an additional job will be submitted, maybe for partitioning?
Is this additional job really necessary?
> CompositeInputFormat in Hama
> ----------------------------
>
> Key: HAMA-774
> URL: https://issues.apache.org/jira/browse/HAMA-774
> Project: Hama
> Issue Type: New Feature
> Components: bsp core
> Affects Versions: 0.6.2
> Reporter: Martin Illecker
> Assignee: Martin Illecker
> Priority: Minor
> Fix For: 0.6.3
>
>
> *Adapting Hadoop CompositeInputFormat to Hama.*
> Adapt the CompositeInputFormat from Hadoop and integrate it into Hama.
> e.g., Useful for matrix multiplication example by doing a inner join of two
> matrices.
> Therefore the Hadoop join package has been adapted within
> *org.apache.hama.bsp.join*.
> Finally I would need some help to integrate the CompositeInputFormat into
> Hama by using
> {code}
> job.setInputFormat(CompositeInputFormat.class);
> job.set("bsp.join.expr", CompositeInputFormat.compose("inner",
> SequenceFileInputFormat.class, aPath, bPath));
> {code}
--
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
