RE: BUG #18195: PL/pgSQL: invalid syntax allowed in SELECT INTO statement

Hi Tom,

Thanks for your answer.
I set plpgsql.extra_errors = 'strict_multi_assignment'

Now, how to explain the following working?

do $sql$
declare
_n int; _s text;
begin
select 1 into _n _s;
raise notice '_n = %, _s = %', _n, _s;
end;
$sql$;

Regards,
Pavel

-----Original Message-----
From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Sent: Tuesday, November 14, 2023 6:31 PM
To: paul(dot)kulakov(at)systematica(dot)ru
Cc: pgsql-bugs(at)lists(dot)postgresql(dot)org
Subject: Re: BUG #18195: PL/pgSQL: invalid syntax allowed in SELECT INTO
statement

PG Bug reporting form <noreply(at)postgresql(dot)org> writes:
> 1. The following code is successfully executed although it has
> incorrect
> syntax: there must be comma (,) between _n and _s in 'into' section.

> select 1, 'string1', 'string2'
> into _n _s;

I believe this is being read the same as

select 1, 'string1', 'string2' _s into _n;

That is, the lack of a comma causes the INTO sub-clause to end, and then _s
is taken as an AS-less column label. As the manual explains, for
backwards-compatibility reasons we allow INTO to be embedded anywhere in the
command, even though that leads to surprising-looking cases like this one.

As for the question of why you don't get an error for the wrong number of
INTO targets, again that's backwards compatibility.
There's a "strict_multi_assignment" check you can turn on to make it
complain about that [1].

regards, tom lane

[1]
https://www.postgresql.org/docs/current/plpgsql-development-tips.html#PLPGSQ
L-EXTRA-CHECKS