Github user HyukjinKwon commented on the issue: https://github.com/apache/spark/pull/19792 I think you can add a new test case somewhere around here - https://github.com/apache/spark/blob/master/python/pyspark/sql/tests.py#L1724 maybe dealing with some combinations of `StructType`, `MapType`, `ArrayType` and one of non-nested type (e.g., `IntegerType`), checking if the error message is really improved. For running tests, you could refer https://spark.apache.org/docs/latest/building-spark.html#pyspark-tests-with-maven. I personally test them locally first and then add and run some tests. Maybe running `./python/run-tests --modules=pyspark-sql` could be enough. Also, make sure running `./dev/lint-python` to check if it complies PEP8 and etc. When you finish them, commit and push which will trigger the build via Jenkins here. See also "Pull Request" in http://spark.apache.org/contributing.html
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org