MIME-Version: 1.0
In-Reply-To: <175987603436.768.10549246436481476867@wrigleys.postgresql.org>
References: <175987603436.768.10549246436481476867@wrigleys.postgresql.org>
From: "David G. Johnston" <david.g.johnston@gmail.com>
Date: Wed, 8 Oct 2025 07:29:06 -0400
Message-ID: 
 <CAKFQuwYHOB1Wp7ZxbHveNrm2PiczNFjy5dPQoJtxT2zJyiU6UA@mail.gmail.com>
Subject: Re: UNION does not append
To: "trescher40@gmx.de" <trescher40@gmx.de>,
	"pgsql-docs@lists.postgresql.org" <pgsql-docs@lists.postgresql.org>
Content-Type: multipart/alternative; boundary="0000000000001be29b0640a3fe1a"
Archived-At: 
 <https://www.postgresql.org/message-id/CAKFQuwYHOB1Wp7ZxbHveNrm2PiczNFjy5dPQoJtxT2zJyiU6UA%40mail.gmail.com>
Precedence: bulk

--0000000000001be29b0640a3fe1a
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Tuesday, October 7, 2025, PG Doc comments form <noreply@postgresql.org>
wrote:

> The following documentation comment has been logged on the website:
>
> Page: https://www.postgresql.org/docs/18/queries-union.html
> Description:
>
> I'm very confused about this statement in the documentation of UNION:
> "UNION
> effectively appends the result of query2 to the result of query1".
> Because, if I append one item to a list, it is still the same list. But i=
n
> the case of UNION, it's ""copied"" into a new anonymous derived-table.
> Alias
> don't work anymore, which I would expect if I just append some rows. Or i=
s
> it like "effectively" means that this statement is not to be taken
> literally? In that case I would really appreciate a more detailed
> description, that after the UNION with tables a new anonymous derived-tab=
le
> is returned and the original first table can't be accessed anymore. Thank
> you for taking your time to read this even if you disagree.
>

Yes, if you place a set-operation using query into a subquery context the
subquery has to be given its own name: and the individual from clause
relations of the contained query are not exposed.  This is just how
subqueries work.

If you don=E2=80=99t place it in a subquery the rows are just sent to the c=
lient.

In both cases the word append simply means =E2=80=9Cusing the column struct=
ure of
the first component query=E2=80=9D.  Though Imsuppose that is strictly =E2=
=80=9Ccolumn
names=E2=80=9D since types can be changed to a more encompassing one if nee=
ded.

David J.

--0000000000001be29b0640a3fe1a
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Tuesday, October 7, 2025, PG Doc comments form &lt;<a href=3D"mailto:nor=
eply@postgresql.org">noreply@postgresql.org</a>&gt; wrote:<br><blockquote c=
lass=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;=
padding-left:1ex">The following documentation comment has been logged on th=
e website:<br>
<br>
Page: <a href=3D"https://www.postgresql.org/docs/18/queries-union.html" tar=
get=3D"_blank">https://www.postgresql.org/<wbr>docs/18/queries-union.html</=
a><br>
Description:<br>
<br>
I&#39;m very confused about this statement in the documentation of UNION: &=
quot;UNION<br>
effectively appends the result of query2 to the result of query1&quot;.<br>
Because, if I append one item to a list, it is still the same list. But in<=
br>
the case of UNION, it&#39;s &quot;&quot;copied&quot;&quot; into a new anony=
mous derived-table. Alias<br>
don&#39;t work anymore, which I would expect if I just append some rows. Or=
 is<br>
it like &quot;effectively&quot; means that this statement is not to be take=
n<br>
literally? In that case I would really appreciate a more detailed<br>
description, that after the UNION with tables a new anonymous derived-table=
<br>
is returned and the original first table can&#39;t be accessed anymore. Tha=
nk<br>
you for taking your time to read this even if you disagree.<br>
</blockquote><div><br></div><div>Yes, if you place a set-operation using qu=
ery into a subquery context the subquery has to be given its own name: and =
the individual from clause relations of the contained query are not exposed=
.=C2=A0 This is just how subqueries work.</div><div><br></div><div>If you d=
on=E2=80=99t place it in a subquery the rows are just sent to the client.</=
div><div><br></div><div>In both cases the word append simply means =E2=80=
=9Cusing the column structure of the first component query=E2=80=9D.=C2=A0 =
Though Imsuppose that is strictly =E2=80=9Ccolumn names=E2=80=9D since type=
s can be changed to a more encompassing one if needed.</div><div><br></div>=
<div>David J.</div><div><br></div>

--0000000000001be29b0640a3fe1a--