On Fri, Oct 08, 2021 at 10:24:33AM -0700, Lily Liu wrote: > 1) Remove DISTINCT > > A simple example is SELECT DISTINCT(name) FROM R. If there is a unique > constraint on the name column. The DISTINCT keyword can be removed safely. > Query plans without the DISTINCT keyword might be much cheaper since > DISTINCT is expensive.
There's an ongoing discussion and patches for this here. Erase the distinctClause if the result is unique by definition https://commitfest.postgresql.org/35/2433/ Perhaps you could help to test or review the patch ? -- Justin