Github user greghogan commented on the pull request:
https://github.com/apache/flink/pull/1956#issuecomment-220147037
I also like @vasia's idea since we might get to add the analytic
`DyadicCensus` (basically counting how many edges are `u <-> v` and `u -> v`).
The relative performance of this implementation will depend on the ratio of
mutual edges. Are there one or more canonical data sets for running a
comparison? The two extremes are an undirected graph and a graph with no mutual
edges, which could be artificially constructed.
Also, I recommend copying the current code into a new class to test a new
implementation.
---
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.
---
