Date: Wed, 29 Apr 2026 13:47:07 +0200
From: Viktor Holmberg <v@viktorh.net>
To: Feike Steenbergen <feikesteenbergen@gmail.com>, David Mullineux
 <dmullx@gmail.com>
Cc: pgsql-performance@postgresql.org
Message-ID: <fffd3a77-f9cb-44e5-91ea-11d6216f3f14@Spark>
In-Reply-To: 
 <CAGsyd8UeMRtmgrQZ1Tg7gMztr=pvF4Dp8mp0kGHzc9ct04FQhw@mail.gmail.com>
References: 
 <CAK_s-G0NrC_KH7kn85arfqkdzvs80GOCCKvz9YbU2=E94qfdPA@mail.gmail.com>
 <CAGsyd8UeMRtmgrQZ1Tg7gMztr=pvF4Dp8mp0kGHzc9ct04FQhw@mail.gmail.com>
Subject: Re: Unexpected Seq Scan's when using MERGE WHEN NOT MATCHED BY
 SOURCE
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="69f1efc3_3a95f874_7e8e"
Archived-At: 
 <https://www.postgresql.org/message-id/fffd3a77-f9cb-44e5-91ea-11d6216f3f14%40Spark>
Precedence: bulk

--69f1efc3_3a95f874_7e8e
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

Apologies for the necromancer here but just chiming in that I also got bi=
tten by this in our production systems today. It=E2=80=99s precisely the =
same pattern - MERGEing into a =E2=80=9Cpart=E2=80=9D of a table. Would b=
e nice if MERGE allowed you to specify a =E2=80=9Cfilter=E2=80=9D or some=
thing so it only affected a subsection of the table - but Toms suggestion=
 of pushing down the condition would also have saved me. =46or now I=E2=80=
=99m just going to have to abandon the MERGE construct and go back to sep=
arate INSERT and DELETE=E2=80=99s.

/Viktor

--69f1efc3_3a95f874_7e8e
Content-Type: text/html; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

<html xmlns=3D=22http://www.w3.org/1999/xhtml=22>
<head>
<title></title>
</head>
<body>
<div name=3D=22messageBodySection=22>
<div dir=3D=22auto=22>Apologies for the necromancer here but just chiming=
 in that I also got bitten by this in our production systems today. It=E2=
=80=99s precisely the same pattern - MERGEing into a =E2=80=9Cpart=E2=80=9D=
 of a table. Would be nice if MERGE allowed you to specify a =E2=80=9Cfil=
ter=E2=80=9D or something so it only affected a subsection of the table -=
 but Toms suggestion of pushing down the condition would also have saved =
me. =46or now I=E2=80=99m just going to have to abandon the MERGE constru=
ct and go back to separate INSERT and DELETE=E2=80=99s.&=23160;</div>
</div>
<div name=3D=22messageSignatureSection=22><br />
<div class=3D=22match=46ont=22>/Viktor</div>
</div>
</body>
</html>

--69f1efc3_3a95f874_7e8e--