Github user cloud-fan commented on the issue:
https://github.com/apache/spark/pull/14864
hi @tejasapatil , I think your second proposal(always preserve bucketing
across all operations) makes sense, but need more discussion to make the
decision. For now, I think it's ok to use your current approach, e.g. only
expose the `outputOrdering` if a bucket only has a single file, but we should
avoid listing files twice.
We can add a lazy val in `FileSourceScanExec`, to store the files after
partition prunning, and use it to calculate `outputOrdering` and `inputRDD`.
What do you think?
---
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 infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
