MIME-Version: 1.0
References: 
 <CAKfDxxzqxm0gS44+4aUDejLq_iszqLJwR+EaM4x2pojrkV58eg@mail.gmail.com>
 <CAKAnmmL327t5MRJ2uonQ3JB5bLegRbfa-WtpaTNBnbJ1iVef5Q@mail.gmail.com>
 <CAKfDxxw-=_naXS-ypU9qWj8eNbAEJvGCy5ZxHA5HW=+M-baDcw@mail.gmail.com>
In-Reply-To: 
 <CAKfDxxw-=_naXS-ypU9qWj8eNbAEJvGCy5ZxHA5HW=+M-baDcw@mail.gmail.com>
From: Jon Zeppieri <zeppieri@gmail.com>
Date: Thu, 19 Dec 2024 14:19:50 -0500
Message-ID: 
 <CAKfDxxwffetLV3UKdLZR_zUOe4odDLges-v7kkTtasfMQ2a2NQ@mail.gmail.com>
Subject: Re: Why a bitmap scan in this case?
To: Greg Sabino Mullane <htamfids@gmail.com>
Cc: pgsql-performance@lists.postgresql.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Archived-At: 
 <https://www.postgresql.org/message-id/CAKfDxxwffetLV3UKdLZR_zUOe4odDLges-v7kkTtasfMQ2a2NQ%40mail.gmail.com>
Precedence: bulk

On Thu, Dec 19, 2024 at 2:09=E2=80=AFPM Jon Zeppieri <zeppieri@gmail.com> w=
rote:
>
> The row estimate is not good. The query estimates 317919 rows but
> there are only 27701. There is some correlation here; if end_on is
> null, start_on is a lot more likely to be recent, so maybe extended
> statistics would be useful here.
>

Though, given that the index only contains rows where end_on is null,
it seems odd that the planner would estimate more rows than are
present in the index. That said, I have no idea whether the planner
uses that sort of information. -J