Github user cloud-fan commented on the issue: https://github.com/apache/spark/pull/20648 > allows the length of tokens are shorter than its schema, putting nulls (or NA) into missing fields Actually I also recalled this is a valid case for csv, and I remember that we did this intentionally. How much can we clean up if we want to keep this behavior in csv? If it's not a lot, maybe we don't need to bother to do refactoring.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org