Github user davies commented on the pull request:
https://github.com/apache/spark/pull/8512#issuecomment-137274773
@piaozhexiu This improvement looks really exciting, great work!
We have some other places that also need this kind of improvements, I'm
learning how to borrow the improvements from here. Basically, the improvements
came from 1) paralleling the listStatus of partitions and 2) using S3 client to
do smart work (we didn't use S3 client in parallel). Is it possible to separate
these two improvements? For example, we could create HadoopRDD for each
partitions in parallel (using thread pool), also using S3 client for each
HadoopRDD (if it's S3), then we have two level of parallel, could be even
faster than current c (in your benchmark). Also, Each partition usually has
single directory, the S3 client will not fetch more files than expected (There
could be some bad case in current approach).
If it works, it's easier to apply these to other places (such as
HadoopFsRelation).
---
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]
