> Exactly. Like Tom I'm not really worried about the proposal, but of
> course I could prove to be wrong. I am ready to assume that bloat in
> pgss entries caused by temp tables is a more common case.
I ran installcheck-parallel with the patch and without the patch
3x, and the reduction in pg_s_s bloat is visible. The same entries
are reused even when the regression database is recreated, as
expected.
## without patch run installcheck-parallel 3 times
postgres=# select count(distinct queryid), count(*) from pg_stat_statements;
count | count
-------+-------
41630 | 74776
(1 row)
## with patch run installcheck-parallel 3 times
postgres=# select count(distinct queryid), count(*) from pg_stat_statements;
count | count
-------+-------
26301 | 73030
(1 row)
>This part of the thread is digressing, but I'd on the side of removing
> entirely the GUC and make the grouping of IN values the default. We
> still have time to discuss that during the beta cycle, so let's do so
> on its related thread.
I will do that tomorrow in that thread. It would not make sense to introduce
a GUC for that behavior and not for this. So, I am glad we discussed this
here.
Looking at the patches, I could not find anything else besides what
was pointed out by Christoph earlier.
--
Sami Imseih
Amazon Web Services (AWS)
