Github user ueshin commented on the issue:
https://github.com/apache/spark/pull/18732
I'm +0 for now.
I'm just wondering whether we can support struct types in vectorized UDF
when needed in the future.
As for adding pandas UDAF, I think we need another decorator or something
to specify it supports partial aggregation or not and the related parameters if
needed.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
