Yep, I've used this with huge success in PostgreSQL when needing to join the contents of tables containing potentially millions of records that could (usually)

Yep, I've used this with huge success in PostgreSQL when needing to join the contents of tables containing potentially millions of records that could (usually) then be filtered via search criteria. Basically I'd do my search-filtering first via a sub-query, then simply join the result of that vs doing it in the opposite order. The performance gains were orders of magnitude.