Github user cloud-fan commented on the issue:
https://github.com/apache/spark/pull/20541
I'm confused about why we need `PushProjectionThroughUnion`. Generally we
only need to push down required columns, not entire project list, as there is
no benifit of doing this. I think we just need to handle `Union` in the
`ColumnPruning` rule, but I may miss something. cc @gatorsmile
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
