Github user viirya commented on the issue:
https://github.com/apache/spark/pull/22357
Thanks! @mallman
For the first query, I think the query plan produced by your WIP patch is
not correct. We don't need to read the `company:struct` from `employer:struct`.
For the second, your WIP patch doesn't push down `IsNotNull(employer)`
predicate into
`FileScan` node.
That is the important difference I noticed for now.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
