MIME-Version: 1.0
References: 
 <CACJPJu-0u221Bker1yn=AJ3xJbW0j29mSM=EvkA_SXecv75hqw@mail.gmail.com>
In-Reply-To: 
 <CACJPJu-0u221Bker1yn=AJ3xJbW0j29mSM=EvkA_SXecv75hqw@mail.gmail.com>
From: Greg Sabino Mullane <htamfids@gmail.com>
Date: Tue, 28 Jan 2025 21:12:03 -0500
Message-ID: 
 <CAKAnmmKQcrF_psSgVy++hUfv02R0h8hgUo7fpnxmO6UFSvEYrQ@mail.gmail.com>
Subject: Re: Simple query with Planner underestimating rows.
To: =?UTF-8?Q?Felipe_L=C3=B3pez_Montes?= <xocas89@gmail.com>
Cc: pgsql-performance@lists.postgresql.org
Content-Type: multipart/alternative; boundary="0000000000000e4b86062cced888"
Archived-At: 
 <https://www.postgresql.org/message-id/CAKAnmmKQcrF_psSgVy%2B%2BhUfv02R0h8hgUo7fpnxmO6UFSvEYrQ%40mail.gmail.com>
Precedence: bulk

--0000000000000e4b86062cced888
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Tue, Jan 28, 2025 at 5:30=E2=80=AFPM Felipe L=C3=B3pez Montes <xocas89@g=
mail.com>
wrote:

> For the sake of clarity and simplicity, I have disabled the nestloop join
> in the session because it involved a gather clause and parallel workers a=
nd
> was still underestimating rows, so the same problem happens with nestloop
> strategy too. Instead now the planner goes for a merge join ( if you stil=
l
> prefer me to send you the nestloop plan, I can do):
>

Yes, please, send the exact plan you are having problems with. Also, what
exactly is the performance issue? It seems your *second best plan* is
running in 36 milliseconds?

Cheers,
Greg

P.S. In the future, you can use this to quickly grab relevant settings:
select name, current_setting(name) from pg_settings where source <>
'default';

--0000000000000e4b86062cced888
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div dir=3D"ltr">On Tue, Jan 28, 2025 at 5:30=E2=80=AFPM F=
elipe L=C3=B3pez Montes &lt;<a href=3D"mailto:xocas89@gmail.com">xocas89@gm=
ail.com</a>&gt; wrote:</div><div class=3D"gmail_quote gmail_quote_container=
"><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;borde=
r-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir=3D"ltr">For th=
e sake of clarity and simplicity, I have disabled the nestloop join in the =
session because it involved a gather clause and parallel workers and was st=
ill underestimating rows, so the same problem happens with nestloop strateg=
y too. Instead now the planner goes for a merge join ( if you still prefer =
me to send you the nestloop plan, I can do):<br></div></blockquote><div><br=
></div><div>Yes, please, send the exact plan you=C2=A0are having problems w=
ith. Also, what exactly is the performance issue? It seems your *second bes=
t plan* is running in 36 milliseconds?</div><div><br></div><div>Cheers,</di=
v><div>Greg</div><div><br></div><div>P.S. In the future, you can use this t=
o quickly grab relevant settings:=C2=A0</div><div>select name, current_sett=
ing(name) from pg_settings where source &lt;&gt; &#39;default&#39;;</div></=
div></div>

--0000000000000e4b86062cced888--