MIME-Version: 1.0
In-Reply-To: <C3E25893-E862-4EA4-825C-BEB5CE3B62C2@gmail.com>
References: <CAEoWx2=mYhCfsnHaN96Qqwq5b0GVS2YgO3zpVqPPRd_iO52wRw@mail.gmail.com>
 <90F9169D-135C-45E5-8221-4F79DAED98E2@gmail.com> <CAKFQuwbWNCX1M1e9_-3P9RVGo10huPjuqdL74FJ+-a5EW791KA@mail.gmail.com>
 <D559C6FA-E101-47F3-BCDF-CEA7FFDF6154@gmail.com> <CAKFQuwZ+o6La5DyK9_-L2gpeiKNvGMbR78UjocO4++JkyZaRHw@mail.gmail.com>
 <A8FA43A6-415E-421A-B3EC-CAB68756975D@gmail.com> <CAKFQuwabxFyFQob=RfKOV3upjxwomAzmGzSouN-h2ypuXz+dZg@mail.gmail.com>
 <46DA7611-C18D-4782-AEFF-F861ECDEFA5C@gmail.com> <CAKFQuwZaquUrTOhfmJ2CenwiyAHx4nORQ0kB01sL+z5pCBuLFQ@mail.gmail.com>
 <CAEoWx2mUiCYJEBuo5D74gi7pHfNz82b54oDHjZNtrXRFDnBPOg@mail.gmail.com>
 <CAKFQuwbQ=L8w8jEaB8_gW-RrNv2_V-VrFYBump7__z-vh-QJWw@mail.gmail.com>
 <CAEoWx2nwoBtWX3=UOQQpGZOuRH5vNv2PHqeFHqDPdiKGaOigiA@mail.gmail.com>
 <CAKFQuwZyJNBN+N5Qtcg0aG6ZpOFvQV5m=cALLgv7g3u1Sz-tZw@mail.gmail.com>
 <245AA9F3-7577-46D6-990C-C308A9F36E82@gmail.com> <CAKFQuwZr_GLBOOLohfEcKVVTcP9D+nzVdUUKabYbXWXRnOFjnA@mail.gmail.com>
 <CAEoWx2=CK=QEvpL5HSgFEuFqjO2Q1sDy6a2G-s-+zv+RzB3mTw@mail.gmail.com>
 <CAN4CZFNoLSUcVKcOOJgOtTkDuOseCL5j9MQr3tGjb3btD=jHNQ@mail.gmail.com>
 <CAKFQuwZwAuO3qXEeqxwV6vM+89BkB-aSkcXqDGG8e_xBy_Q3Xw@mail.gmail.com> <C3E25893-E862-4EA4-825C-BEB5CE3B62C2@gmail.com>
From: "David G. Johnston" <david.g.johnston@gmail.com>
Date: Sun, 25 Jan 2026 21:36:15 -0700
Message-ID: <CAKFQuwa9OEJ_sygQi3fNKxzQz7M6V7-zu=4fL0rquXCjudUoyw@mail.gmail.com>
Subject: Re: docs: clarify ALTER TABLE behavior on partitioned tables
To: Chao Li <li.evan.chao@gmail.com>
Cc: Postgres hackers <pgsql-hackers@lists.postgresql.org>, 
	Zsolt Parragi <zsolt.parragi@percona.com>, Amit Kapila <amit.kapila16@gmail.com>
Content-Type: multipart/alternative; boundary="0000000000003123a00649430ca4"
Archived-At: <https://www.postgresql.org/message-id/CAKFQuwa9OEJ_sygQi3fNKxzQz7M6V7-zu%3D4fL0rquXCjudUoyw%40mail.gmail.com>
Precedence: bulk

--0000000000003123a00649430ca4
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Sunday, January 25, 2026, Chao Li <li.evan.chao@gmail.com> wrote:

>
> >      <para>
> >       When applied to a partitioned table, partition columns and
> constraints
> >       are implicitly renamed.
> >       Specifying <literal>ONLY</literal> is not allowed, and this comma=
nd
> >       cannot be used on individual partitions.
> >      </para>
>
> But for "Specifying <literal>ONLY</literal> is not allowed, and this
> command, cannot be used on individual partitions.=E2=80=9D, that doesn=E2=
=80=99t seem
> correct. See my test:
> ```
> evantest=3D# create table root (i int, j int) partition by list(i);
> CREATE TABLE
> evantest=3D# create table p1 partition of root for values in (1);
> CREATE TABLE
> evantest=3D# alter table p1 rename to pp1; <=3D=3D Rename a partition is =
allowed.
> ALTER TABLE
> evantest=3D# alter table only pp1 rename to p1; <=3D=3D ONLY can be used,=
 but
> just no effect
> ALTER TABLE
> ```
>

I was mentally restricting the second sentence about ONLY to the column and
constraints renaming action, which are called out by the first sentence.
It makes little sense to talk about renaming the table, parent or child,
here in the context of ONLY.  It goes without mention that table renaming
never cascades.  Only is implied for that action, even if only should just
be considered valid on a parent in any case.

But I=E2=80=99d accept a sentence like: =E2=80=9CTable renames always only =
apply to the
named table.=E2=80=9D Added to that paragraph; it=E2=80=99s a convoluted co=
mmand.

David J.

--0000000000003123a00649430ca4
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Sunday, January 25, 2026, Chao Li &lt;<a href=3D"mailto:li.evan.chao@gma=
il.com">li.evan.chao@gmail.com</a>&gt; wrote:<br><blockquote class=3D"gmail=
_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:=
1ex"><br>&gt;=C2=A0 =C2=A0 =C2=A0 &lt;para&gt;<br>
&gt;=C2=A0 =C2=A0 =C2=A0 =C2=A0When applied to a partitioned table, partiti=
on columns and constraints<br>
&gt;=C2=A0 =C2=A0 =C2=A0 =C2=A0are implicitly renamed.<br>
&gt;=C2=A0 =C2=A0 =C2=A0 =C2=A0Specifying &lt;literal&gt;ONLY&lt;/literal&g=
t; is not allowed, and this command<br>
&gt;=C2=A0 =C2=A0 =C2=A0 =C2=A0cannot be used on individual partitions.<br>
&gt;=C2=A0 =C2=A0 =C2=A0 &lt;/para&gt;<br><br>
But for &quot;Specifying &lt;literal&gt;ONLY&lt;/literal&gt; is not allowed=
, and this command, cannot be used on individual partitions.=E2=80=9D, that=
 doesn=E2=80=99t seem correct. See my test:<br>
```<br>
evantest=3D# create table root (i int, j int) partition by list(i);<br>
CREATE TABLE<br>
evantest=3D# create table p1 partition of root for values in (1);<br>
CREATE TABLE<br>
evantest=3D# alter table p1 rename to pp1; &lt;=3D=3D Rename a partition is=
 allowed.<br>
ALTER TABLE<br>
evantest=3D# alter table only pp1 rename to p1; &lt;=3D=3D ONLY can be used=
, but just no effect<br>
ALTER TABLE<br>
```<br>
</blockquote><div><br></div><div>I was mentally restricting the second sent=
ence about ONLY to the column and constraints renaming action, which are ca=
lled out by the first sentence.=C2=A0 It makes little sense to talk about r=
enaming the table, parent or child, here in the context of ONLY.=C2=A0 It g=
oes without mention that table renaming never cascades.=C2=A0 Only is impli=
ed for that action, even if only should just be considered valid on a paren=
t in any case.</div><div><br></div><div>But I=E2=80=99d accept a sentence l=
ike: =E2=80=9CTable renames always only apply to the named table.=E2=80=9D =
Added to that paragraph; it=E2=80=99s a convoluted command.</div><div><br><=
/div><div>David J.</div><div><br></div>

--0000000000003123a00649430ca4--