Github user andrewor14 commented on the pull request:
https://github.com/apache/spark/pull/4878#issuecomment-77522081
I wasn't expecting this to be closed. Although parts of the changes didn't
seem particularly useful, I believe the changes in `ExternalSorter` could
potentially lead to a performance gain. There we open a single output stream to
merge multiple partition files without doing any sort of buffering, which is
pretty inefficient if we have many such files. @jerryshao's comment about NIO's
`transferTo` being used refers to the read path, not the write path, so doing
buffering on the latter may actually matter.
@ravipesala I think there is some value in doing some sort of performance
benchmarking before and after this change before deciding once and for all that
it's not worth it.
---
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]
