Consider this scenario:
You have a group of PFs, which in turn have a few LFs on them. You are
mainly doing native/RLA processing on them, but also some SQL, both
views and ad-hoc queries, all of which can potentially involve SQL
joins, and most of which involve programmatically-created SQL.
So far, so good.
But now, let's change the scenario:
Suppose you want to replace the PFs and LFs with join-logicals (or SQL
join-views), tying directly into data in other libraries.
But what about the aforementioned SQL views and queries, that involve
joins? Do they have to be built from the raw PFs? Or can they be built
on top of the existing joins?