Content-Type: text/plain;
	charset=utf-8
Mime-Version: 1.0 (Mac OS X Mail 16.0 \(3864.600.51.1.1\))
Subject: Re: doc: Clarify ALTER CONSTRAINT enforceability wording
From: Chao Li <li.evan.chao@gmail.com>
In-Reply-To: <711B1ED3-1781-4B6C-A573-B58AF20770E5@gmail.com>
Date: Tue, 26 May 2026 13:42:48 +0800
Cc: Andrew Dunstan <andrew@dunslane.net>,
 jian he <jian.universality@gmail.com>
Content-Transfer-Encoding: quoted-printable
Message-Id: <72669A61-01A6-4B6C-A9A7-46BDF094B875@gmail.com>
References: <711B1ED3-1781-4B6C-A573-B58AF20770E5@gmail.com>
To: PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Archived-At: 
 <https://www.postgresql.org/message-id/72669A61-01A6-4B6C-A9A7-46BDF094B875%40gmail.com>
Precedence: bulk


> On May 25, 2026, at 17:31, Chao Li <li.evan.chao@gmail.com> wrote:
>=20
> Hi,
>=20
> I just started to test =E2=80=9CAdd support for altering CHECK =
constraint enforceability=E2=80=9D. I read the related ALTER TABLE doc =
first, and found it a bit confusing:
> ```
> ALTER CONSTRAINT=20
>    This form alters the attributes of a constraint that was previously =
created. Currently FOREIGN KEY and CHECK constraints may be altered in =
this fashion, but see below.
> ```
>=20
> In the feature commit 342051d73, only =E2=80=9CCHECK" was added:
> ```
> --- a/doc/src/sgml/ref/alter_table.sgml
> +++ b/doc/src/sgml/ref/alter_table.sgml
> @@ -578,8 +578,8 @@ WITH ( MODULUS <replaceable =
class=3D"parameter">numeric_literal</replaceable>, REM
>     <listitem>
>      <para>
>       This form alters the attributes of a constraint that was =
previously
> -      created. Currently only foreign key constraints may be altered =
in
> -      this fashion, but see below.
> +      created. Currently <literal>FOREIGN KEY</literal> and =
<literal>CHECK</literal>
> +      constraints may be altered in this fashion, but see below.
>      </para>
> ```
>=20
> However, the last phrase =E2=80=9Cbut see below=E2=80=9D is quite =
confusing. I read the doc several times up and down, and finally my best =
guess is that =E2=80=9Csee below=E2=80=9D refers to the immediately =
following paragraph=E2=80=99s sentence =E2=80=9COnly not-null =
constraints may be altered in this fashion at present=E2=80=9D, but the =
sentence does not say anything about enforceability.
> ```
> ALTER CONSTRAINT ... INHERIT
> ALTER CONSTRAINT ... NO INHERIT=20
>      These forms modify an inheritable constraint so that it becomes =
not inheritable, or vice-versa. **Only not-null constraints may be =
altered in this fashion at present.** In addition to changing the =
inheritability status of the constraint, in the case where a =
non-inheritable constraint is being marked inheritable, if the table has =
children, an equivalent constraint will be added to them. If marking an =
inheritable constraint as non-inheritable on a table with children, then =
the corresponding constraint on children will be marked as no longer =
inherited, but not removed.
> ```
>=20
> As I understand it, the current behavior is:
> * Deferrability attributes can only be altered for FOREIGN KEY =
constraints
> * Enforceability attributes can be altered for both FOREIGN KEY and =
CHECK constraints, and CHECK support was newly added by 342051d73 for =
v19
>=20
> To remove the confusion, I updated the documentation. See the attached =
patch for details.
>=20
> Best regards,
> --
> Chao Li (Evan)
> HighGo Software Co., Ltd.
> https://www.highgo.com/
>=20
>=20
>=20
>=20
> <v1-0001-doc-Clarify-ALTER-CONSTRAINT-enforceability-behav.patch>

As I found a bug of the feature and proposed a fix in [1], let's merge =
this thread into that one, as they are for the same feature.

[1] https://postgr.es/m/E74C57FA-1DD0-4C8E-8FB1-538034752592@gmail.com

Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/