If I have a Spark Dataset, I can do the following operations :

• Map, for a 1-to-1 relationship from T to U types
• FlatMap, for a 1-to-many relationship from T to U types
• Reduce, for a all-to-1 relationship from T types to one T.

What about the following two relationships ?

• many-to-1 relationship from T types to a U type.
• many-to-many relationship from T types to U types ?

I don’t see anything in the Dataset documentation. Am I supposed to do a GroupBy then a Map/FlatMap ?