Added missing rdd.distinct(numPartitions) and associated tests

Author: Matthew Farrellee <matt@redhat.com>

Closes #2383 from mattf/SPARK-3519 and squashes the following commits:

30b837a [Matthew Farrellee] Combine test cases to save on JVM startups
6bc4a2c [Matthew Farrellee] [SPARK-3519] add distinct(n) to SchemaRDD in PySpark
7a17f2b [Matthew Farrellee] [SPARK-3519] add distinct(n) to PySpark