Github user cloud-fan commented on the issue: https://github.com/apache/spark/pull/23171 I think `InSet` is not an optimized version of `In`, but just a way to separate the implementation for different conditions (the length of the list). Maybe we should do the same thing here, create a `InSwitch` and convert `In` to it when meeting some conditions. One problem is, `In` and `InSwitch` is same in the interpreted version, maybe we should create a base class for them.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org