Github user mateiz commented on the pull request:
https://github.com/apache/spark/pull/2524#issuecomment-63863153
BTW when doing this only for result stages, my suggestion is to use the
data structures within the stage instead of having a second HashMap. I believe
I mentioned this before too (maybe on the previous PR): all you need to do is
move the accumulator update code within the `if (!job.finished(rt.outputId)) {`
for such stages, similar to how it only fetches results once for each task.
Again the point is to avoid adding a new data structure in DAGScheduler that we
must then carefully manage and clean up.
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
