Hi, my release is 9.5.4. a took a look over it. I guessed that counting could be slow because it needs to read everything and also that it can take advantage from an index. But I don't understand why the delay is after the updates for a
Best guess, autovacuum kicked in and marked a bunch of rows as no longer in play and thereby reduced the number of rows that needed to be counted.
certain time and why WHERE..IN is much faster (ok, it's an index, but I'm reading all the rows).
So per the second link have you tried something like:
SELECT COUNT(*) FROM Table WHERE id > 0;
Hi,
no I
did not (yet). But I guess that it would be similar to the one or to the other. I will give updates if I try.
