Appreciate the clarifications. The only thing I would then add is that if we agree that it's unintuitive, this could be better reflected in the documentation.
Currently the only mention is the following (pgstatstatements.html):
"queries with identical texts might appear as separate entries, if they have different meanings as a result of factors such as different search_path settings."
Thanks and best regards,
Marcus Kempe
On Fri, Oct 28, 2022 at 3:28 AM Julien Rouhaud <rjuju123@gmail.com> wrote:
On Thu, Oct 27, 2022 at 02:28:17PM -0400, Tom Lane wrote: > Julien Rouhaud <rjuju123@gmail.com> writes: > > On Thu, Oct 27, 2022 at 01:49:18PM +0000, PG Bug reporting form wrote: > >> I would have expected both calls to be normalized and fingerprinted to one > >> and the same queryId. > > > Agreed, and that's actually a known problem that is currently being worked on. > > You can look at > > https://www.postgresql.org/message-id/flat/36e5bffe-e989-194f-85c8-06e7bc88e6f7%40amazon.com > > for details about the discussion and in-progress fix. > > Um ... that seems unrelated. AFAICS the OP is complaining about > the fact that '42'::int4::int8 is not identical to '42'::int8. > Well, they're not. I seriously doubt that we would or should > consider trying to get queryjumble to mask that.
Oh wow I totally misread the bug report, sorry about that. A good night sleep and a coffee later I see what it's about, and I agree that we shouldn't try to mask that even if it might seem surprising.