Github user HyukjinKwon commented on a diff in the pull request: https://github.com/apache/spark/pull/20211#discussion_r160860320 --- Diff: python/pyspark/sql/group.py --- @@ -233,6 +233,27 @@ def apply(self, udf): | 2| 1.1094003924504583| +---+-------------------+ + Notes on grouping column: --- End diff -- Yup, I saw this usecase as described in the JIRA and I got that the specific case can be simplified; however, I am not sure if it's straightforward to the end users. For example, if I use `pandas_udf` I think I would simply expect the return schema is matched as described in `returnType`. I think `pandas_udf` already need some background and I think we should make it simpler as possible as we can. It might be convenient to make the guarantee on grouping columns in some cases vs this might be a kind of magic inside. I would prefer to let the UDF to specify the grouping columns to make this more straightforward more ..
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org