Major Description
One of the pre-requisite for the feature is to allow V2 input partitions to report their partition values to Spark, which can use them to compare if both sides of join are co-partitioned, and also optionally group input partitions together.
