Comment by jellyfishbeaver

We actually baked a rule to catch UDF usage into our Python linter. Almost always, a UDF can be refactored to use only native PySpark functions.