Github user viirya commented on the pull request:
https://github.com/apache/spark/pull/10539#issuecomment-170811672
@ankurdave @jkbradley I tried to replace calls to
`GraphImpl.fromExistingRDDs(vertices, edges)` with `Graph(vertices, edges)` as
@ankurdave suggested. And the two tests fail.
I did some experiments and it is interesting to me that the following has
different results:
val normalized = GraphImpl.fromExistingRDDs(vD, gA.edges)
.mapTriplets(
e => e.attr / math.max(e.srcAttr, MLUtils.EPSILON),
TripletFields.Src)
val normalized2 = Graph(normalized.vertices, normalized.edges)
If I return `normalized` from the method `normalize`, the first test
succeeds. But if I return `normalized2`, the test fails. Because I create new
graph with the vertices and edges from the same graph `normalized`, I suppose
that `normalized2` should be as same as `normalized`. Routing tables should not
be the problem here because the edges and vertices are from the same graph, so
the routing table should be correct.
So I am interested in why `normalized` succeeds and `normalized2` fails.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
