Github user HyukjinKwon commented on the issue:
https://github.com/apache/spark/pull/14660
@rxin, I thought reading a small file is a possible corner case. In this
case, it would not be only a small fraction. Improving it without a regression
might be a legitimate optimization by preventing to launch another job.
It is true that in such (or most cases) users would just tolerate the
elapsed time but I guess we have added such improvements time to time to
improve the performance in general bit by bit.
`spark.sql.sources.parallelPartitionDiscovery.threshold` might be the
option for the similar cases.
I guess this PR does not introduce another a bunch of newly added codes to
maintain more but just use the original codes.
---
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]
