Github user heary-cao commented on the issue:
https://github.com/apache/spark/pull/20541
in my opinion, this is considered that PushProjectionThroughUnion optimizes
rules when there are multiple columns of union in data sources, while
projection requires only a few columns, and the performance of file operation
is better. thanks.--- --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
