On Sun, Jul 26, 2020 at 11:34 AM Tomas Vondra <tomas.von...@2ndquadrant.com> wrote: > That's 1.6GB, if I read it right. Which is more than 200MB ;-)
Sigh. That solves that "mystery": the behavior that my sorted vs random example exhibited is a known limitation in hash aggs that spill (and an acceptable one). The memory usage is reported on accurately by EXPLAIN ANALYZE. -- Peter Geoghegan
