dongjoon-hyun edited a comment on pull request #28833: URL: https://github.com/apache/spark/pull/28833#issuecomment-655291242
First of all, it's not a good idea to add `NullType` into a new Spark datatype officially. Not only the exposure causes more complexity, but also what can we do in Spark SQL world (https://spark.apache.org/docs/latest/sql-ref-datatypes.html) if that is an official type? > If we'll still care and have NullType, let's make it a proper type in Spark. Previously, this was supported until Apache Spark 2.0.0. After that, Apache Spark didn't support void. This PR also tried to forbid `VOID`. `AstBuilder` provides a way for graceful warning. Currently, we are very careful even in the error message, we didn't mention `void type`. We called it `unknown type`. I believe this PR is one way to implement your idea, too. Of course, we can add more messages, too. > If we're going to treat void as Hive legacy, let's don't support it at all and make the direction to deprecate and remove NullType away. In any way, since this is a legitimate suggestion from @HyukjinKwon , cc @gatorsmile , too. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org