Github user zhichao-li commented on the pull request:
https://github.com/apache/spark/pull/9097#issuecomment-153934785
@chenghao-intel Just tested with data which have 15w small files and 1000
partitions.
1) SQL (select count(*) from test), only improve a little bit, I guess
tasks scheduling is not the bottle neck here, so reducing tasks number would
not have too much effect.
2) SQL (select count(*) from test group by a ), the performance would
increase by 3 times. reducing the tasks would largely improve the shuffle
performance.
---
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]
