Hi folks,
Apologies if it's the wrong place to ask. I have a question for postgres
developers.
I wrote parser for COPY binary encoding. Everything is great but one thing
worries me. It seems like the time offset field has reversed digit sign.
I'm referring to this function
https://github.com/pos
oblem more repeatedly? I just want to write a regression test to
make sure it doesn't happen again.
--
Best regards, Ruslan.
On Tue, Jun 13, 2023 at 1:26 PM Thomas Kellerer wrote:
> Ruslan Zakirov schrieb am 13.06.2023 um 09:49:
> > For example I have a query:
> >
> > SELECT main.*, count(*) OVER () FROM (SELECT DISTINCT ... ORDER BY X)
> > main;
> >
> > So the `ORDER BY` cla
On Tue, Jun 13, 2023 at 6:06 PM Tom Lane wrote:
> Ruslan Zakirov writes:
> > I know how to fix the problem and I know that ORDER BY should be in the
> > outermost select.
>
> > However, I want to write a test case that shows that the old code is
> wrong,
> > b
On Tue, Jun 13, 2023 at 8:55 PM Ruslan Zakirov
wrote:
> On Tue, Jun 13, 2023 at 6:06 PM Tom Lane wrote:
>
>> Ruslan Zakirov writes:
>> > I know how to fix the problem and I know that ORDER BY should be in the
>> > outermost select.
>>
>> > However
umentation. No prove in tests that the old code
was wrong
and a new one fixes it.
> the subquery *unless* one forces a different order on the
> outer query. Which in turn would defeat the purpose as then
> the outer query *does* have an explicit ordering...
>
--
Best regards, Ruslan.
can try the following index:
CREATE INDEX conversation_item_ruz1 ON conversation_item(item_uuid,
conversation_uuid, tenant_id);
I believe this index would allow Pg to use "index only scan" as variation
of Plan A and avoid touching the conversation_item table completely.
--
Best regards, Ruslan.
