Github user HyukjinKwon commented on a diff in the pull request: https://github.com/apache/spark/pull/20798#discussion_r175332547 --- Diff: docs/sql-programming-guide.md --- @@ -1799,6 +1799,12 @@ different than a Pandas timestamp. It is recommended to use Pandas time series f working with timestamps in `pandas_udf`s to get the best performance, see [here](https://pandas.pydata.org/pandas-docs/stable/timeseries.html) for details. +### Keyword Arguments and Related Structures + +Currently, for `pandas_udf` it is not possible to pass keyword arguments to a function. The wrapped +function must also not be a `functools.partial` function object. Functions with a zero-length argument +list are unsupported, but can be approximated via a single-argument udf which ignores the passed arg. --- End diff -- Ah, hm .. I think we are fine to leave this note out of the SQL programing guide for now .. Arguably this seems rather a corner case (just given my monitoring mailing list and JIRAs so far). I personally have promoted guys to leave some notes about key points only.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org