[
https://issues.apache.org/jira/browse/HAMA-790?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13773722#comment-13773722
]
Edward J. Yoon commented on HAMA-790:
-------------------------------------
Sorry for late reviewing... I was on vacation. The failure of your unit test
program is related with input data structure and vertex reader.
The current Graph framework assumes that each record includes a vertex and its
edges.
{code}
while ((record = peer.readNext()) != null) {
converted = converter.convertRecord(record, conf);
vertex = (Vertex<V, E, M>) converted.getKey();
vertex.runner = this;
vertex.setup(conf);
if (selfReference) {
vertex.addEdge(new Edge<V, E>(vertex.getVertexID(), null));
}
vertices.addVertex(vertex);
// Reinitializing vertex object for memory based implementations of
// VerticesInfo
vertex = GraphJobRunner.<V, E, M> newVertexInstance(VERTEX_CLASS);
vertex.runner = this;
}
{code}
So, you should generate the input data, like below:
v1 - edges: e1, e2, e3
instead of:
v1 - edges: e1
v1 - edges: e2
v1 - edges: e3
...
However, I don't think there's a logical problem in your program. You can solve
the issue by fixing the framework.
> Create a unit test for the SemiClustering example
> -------------------------------------------------
>
> Key: HAMA-790
> URL: https://issues.apache.org/jira/browse/HAMA-790
> Project: Hama
> Issue Type: Sub-task
> Components: graph, test
> Reporter: Edward J. Yoon
> Assignee: Renil J
> Attachments: HAMA-790-V1.patch
>
>
> Please create a unit test for the SemiClustering example.
--
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
