From: Tom Lane <tgl@sss.pgh.pa.us>
To: Logan Grosz <logan.grosz@gmail.com>
cc: pgsql-general@lists.postgresql.org
Subject: Re: Is it correct to raise an exception in a domain check
In-reply-to: <CANn=KGi1fU3rkEbwnOZX3_WNQJXa39tbkmn7N4tAMSf75r7nNw@mail.gmail.com>
References: <CANn=KGi1fU3rkEbwnOZX3_WNQJXa39tbkmn7N4tAMSf75r7nNw@mail.gmail.com>
Comments: In-reply-to Logan Grosz <logan.grosz@gmail.com>
	message dated "Sun, 08 Jun 2025 22:01:47 -0600"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <805458.1749591064.1@sss.pgh.pa.us>
Date: Tue, 10 Jun 2025 17:31:04 -0400
Message-ID: <805459.1749591064@sss.pgh.pa.us>
Archived-At: <https://www.postgresql.org/message-id/805459.1749591064%40sss.pgh.pa.us>
Precedence: bulk

Logan Grosz <logan.grosz@gmail.com> writes:
> I have a `DOMAIN` on `JSONB`. I need to verify the shape of the JSON, so I
> have a check expression written in PL/pgSQL. The docs say

>> Each constraint must be an expression producing a Boolean result

> Would it be correct to raise an exception to indicate a failed
> check?

It'd be better just to have the CHECK expression return false.
Admittedly, that's usually just going to end in an exception,
but pre-judging that inside the expression doesn't seem ideal.
An example of why not is that you'd break "soft" input error
handling in COPY.

As you say, there are some cases where it's hard to avoid
an exception, but I'm not sure that "better error message"
is a good justification for throwing one.  Still, in the
end it's your own judgment to make.

			regards, tom lane