[
https://issues.apache.org/jira/browse/HAMA-737?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13589416#comment-13589416
]
Hudson commented on HAMA-737:
-----------------------------
Integrated in Hama trunk #108 (See
[https://builds.apache.org/job/Hama%20trunk/108/])
[HAMA-737]: Add graph output writer (Revision 1451131)
Result = SUCCESS
tjungblut :
Files :
* /hama/trunk/CHANGES.txt
*
/hama/trunk/graph/src/main/java/org/apache/hama/graph/DefaultVertexOutputWriter.java
* /hama/trunk/graph/src/main/java/org/apache/hama/graph/GraphJob.java
* /hama/trunk/graph/src/main/java/org/apache/hama/graph/GraphJobRunner.java
* /hama/trunk/graph/src/main/java/org/apache/hama/graph/VertexOutputWriter.java
> Add graph output writer
> -----------------------
>
> Key: HAMA-737
> URL: https://issues.apache.org/jira/browse/HAMA-737
> Project: Hama
> Issue Type: New Feature
> Components: graph
> Affects Versions: 0.6.0
> Reporter: Thomas Jungblut
> Assignee: Thomas Jungblut
> Fix For: 0.6.1, 0.7.0
>
> Attachments: HAMA-737_1.patch
>
>
> Currently only VertexID\tVertexValue can be written to the output.
> If someone want to write anything else (e.G. the edge values) he can not.
> Analogously to the VertexInputReader, it would be nice to have a
> VertexOutputWriter.
> It should have the following abilities:
> - setup (Configuration)
> - write(Vertex<>, Peer)
> The current state would be mimicked by a DefaultVertexOutputWriter that
> writes the vertex key and value to the peer.
> The current writing happens in cleanup of the GraphJobRunner.
> Following things must be added:
> - A reflection based instantiation process that reads a class from
> Configuration and makes it available through the VertexOutputWriter interface
> in the GraphJobRunner.
> - cleanup must be calling the #write method for each vertex
> - GraphJob should have another method called setVertexOutputWriterClass(class)
> - GraphJob must ensure in submit override that the DefaultVertexOutputWriter
> is set if nothing was configured yet.
> Usecase:
> Actually every algorithm that changes edge values and use it for further post
> processing. Also algorithms that need to hold state in the vertex that is not
> the value, currently this can only be solved by a multiple valued Writable
> Vertex value. (Bipartite Matching could be simplified in that regard).
--
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
