[
https://issues.apache.org/jira/browse/GIRAPH-228?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13405099#comment-13405099
]
Eli Reisman commented on GIRAPH-228:
------------------------------------
Hi Avery,
Stupid question (as usual!) but the ArrayListWritable type looks perfect and
easy to use too. But, its edu.umb.cloud9 package would need to be added to the
maven dependencies for Giraph and it has other things it depends on that now
also come along for the ride as I understand it. Is that OK to add if I
re-convert my example vertex to use it, or is this why is has not been included
so far? Would it be better to suggest its use to multigraph implementors as a
self-downloaded jar and forego an example vertex or is it OK to include this
dependency change in a patch on JIRA?
If it came down to just pointing people in the right direction, that would be
great right there as the first things that come up when trying to figure out
how to support arrays is all the hadoop.io stuff that doesn't seem to behave
well in Giraph vertices.
The patch I have posted here is to change GIRAPH-217
(SimpleTriangleClosingVertex) into a form that uses Text for final output, and
adds a test or two. So its probably a worthwhile improvement over 217 either
way. But if I could get the array types into the Giraph POM so others could run
the example, I'd be happy to skip this patch and upload another fix using them.
I do think there's value in having an array-based example in the examples code.
Thanks for your advice.
> SimpleTriangleClosingVertex should not use ArrayWritable for a vertex value
> ---------------------------------------------------------------------------
>
> Key: GIRAPH-228
> URL: https://issues.apache.org/jira/browse/GIRAPH-228
> Project: Giraph
> Issue Type: Bug
> Components: examples, test
> Affects Versions: 0.2.0
> Reporter: Eli Reisman
> Assignee: Eli Reisman
> Priority: Trivial
> Fix For: 0.2.0
>
> Attachments: GIRAPH-228-1.patch
>
>
> Need to convert to (yuck) Text output for vertex values, hope my tuples don't
> get too long. I followed several instructions from Hadoop about subclassing
> and using existing ArrayWritables, always both methods will compile with 'mvn
> verify' nicely, but break in actual use. GiraphJob.run() seems to be the
> culprit. Don't know if I can solve this myself in a way that will be pleasing
> to everyone. See GIRAPH-227 if you can help fix this ArrayWritable problem.
> As far as the vertex example, I'll covert it (and related tests) to Text
> output now.
> Thanks again to Sean Choi, who noticed this and used the same (Text for
> tuples) solution successfully. ;)
>
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
