Github user cloud-fan commented on the issue:
https://github.com/apache/spark/pull/18732
> Use different function name for different input/output type
Yea it's a bad idea as there are many combinations, and I just wanna use
different APIs for different scenarios, e,g, `@pandas_udf` for
select/withColumn(`Series* -> Series`), `@pandas_grouped_udf` for
groupBy(apply: `DataFrame -> DataFrame`, reduce: `DataFrame -> Scalar` and
more) and `@pandas_udaf` for aggregate.
Different scenarios usually have different requirements, having different
APIs can help us satisfy these requirements individually.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
