BUG #18693: Column names not set when using SELECT STRICT INTO with RECORD type
The following bug has been logged on the website:
Bug reference: 18693
Logged by: Jan Behrens
Email address: jbe-mlist@magnetkern.de
PostgreSQL version: 16.4
Operating system: FreeBSD
Description:
I used the following SQL code:
CREATE TABLE tbl (a TEXT, b TEXT);
CREATE FUNCTION "buggy"() RETURNS VOID LANGUAGE plpgsql AS $$
DECLARE
t tbl%ROWTYPE;
r RECORD;
BEGIN
SELECT a, b INTO t FROM (SELECT 'A' AS a, 'B' AS b) AS q;
RAISE NOTICE 'non-strict t = %', t;
RAISE NOTICE 'non-strict t.a = %', t.a;
RAISE NOTICE 'non-strict t.b = %', t.b;
SELECT a, b STRICT INTO t FROM (SELECT 'A' AS a, 'B' AS b) AS q;
RAISE NOTICE 'STRICT t = %', t;
RAISE NOTICE 'STRICT t.a = %', t.a;
RAISE NOTICE 'STRICT t.b = %', t.b;
SELECT a, b INTO r FROM (SELECT 'A' AS a, 'B' AS b) AS q;
RAISE NOTICE 'non-strict r = %', r;
RAISE NOTICE 'non-strict r.a = %', r.a;
RAISE NOTICE 'non-strict r.b = %', r.b;
SELECT a, b STRICT INTO r FROM (SELECT 'A' AS a, 'B' AS b) AS q;
RAISE NOTICE 'STRICT r = %', r;
RAISE NOTICE 'STRICT r.a = %', r.a;
RAISE NOTICE 'STRICT r.b = %', r.b;
END;
$$;
SELECT buggy();
Invoking the buggy() function resulted in:
NOTICE: non-strict t = (A,B)
NOTICE: non-strict t.a = A
NOTICE: non-strict t.b = B
NOTICE: STRICT t = (A,B)
NOTICE: STRICT t.a = A
NOTICE: STRICT t.b = B
NOTICE: non-strict r = (A,B)
NOTICE: non-strict r.a = A
NOTICE: non-strict r.b = B
NOTICE: STRICT r = (A,B)
NOTICE: STRICT r.a = A
ERROR: record "r" has no field "b"
CONTEXT: SQL expression "r.b"
PL/pgSQL function buggy() line 21 at RAISE
I would expect the last statement to not result in an error but to correctly
access field "b" as in the non-strict case or tbl%ROWTYPE case.
PG Bug reporting form <noreply@postgresql.org> writes:
I used the following SQL code:
CREATE TABLE tbl (a TEXT, b TEXT);
CREATE FUNCTION "buggy"() RETURNS VOID LANGUAGE plpgsql AS $$
DECLARE
t tbl%ROWTYPE;
r RECORD;
BEGIN
SELECT a, b INTO t FROM (SELECT 'A' AS a, 'B' AS b) AS q;
RAISE NOTICE 'non-strict t = %', t;
RAISE NOTICE 'non-strict t.a = %', t.a;
RAISE NOTICE 'non-strict t.b = %', t.b;
SELECT a, b STRICT INTO t FROM (SELECT 'A' AS a, 'B' AS b) AS q;
RAISE NOTICE 'STRICT t = %', t;
RAISE NOTICE 'STRICT t.a = %', t.a;
RAISE NOTICE 'STRICT t.b = %', t.b;
SELECT a, b INTO r FROM (SELECT 'A' AS a, 'B' AS b) AS q;
RAISE NOTICE 'non-strict r = %', r;
RAISE NOTICE 'non-strict r.a = %', r.a;
RAISE NOTICE 'non-strict r.b = %', r.b;
SELECT a, b STRICT INTO r FROM (SELECT 'A' AS a, 'B' AS b) AS q;
RAISE NOTICE 'STRICT r = %', r;
RAISE NOTICE 'STRICT r.a = %', r.a;
RAISE NOTICE 'STRICT r.b = %', r.b;
END;
$$;
You have the syntax wrong: what you intended (I presume) is
SELECT a, b INTO STRICT r FROM (SELECT 'A' AS a, 'B' AS b) AS q;
With what you wrote, the STRICT is taken as an output column alias of
the outer SELECT, as if the command were
SELECT a, b AS "strict" INTO r FROM (SELECT 'A' AS a, 'B' AS b) AS q;
The alias makes no difference when assigning to "t", since its column
names are predetermined. But when assigning to "r", the second column
winds up being named "strict" not "b". You can demonstrate this with
SELECT a, b STRICT INTO r FROM (SELECT 'A' AS a, 'B' AS b) AS q;
RAISE NOTICE 'STRICT r = %', r;
RAISE NOTICE 'STRICT r.a = %', r.a;
RAISE NOTICE 'STRICT r.strict = %', r."strict";
(It's kind of annoying that "strict" has to be double-quoted
in the RAISE NOTICE, especially since you get a rather misleading
error if it isn't. But that seems like a different discussion.)
regards, tom lane
On Wed, 06 Nov 2024 17:30:02 -0500
Tom Lane <tgl@sss.pgh.pa.us> wrote:
PG Bug reporting form <noreply@postgresql.org> writes:
[...]You have the syntax wrong: what you intended (I presume) is
SELECT a, b INTO STRICT r FROM (SELECT 'A' AS a, 'B' AS b) AS q;
With what you wrote, the STRICT is taken as an output column alias of
the outer SELECT, [...]regards, tom lane
Oooops, my apologies! It totally makes sense now.
Thanks for your thorough/helpful response.
Regards
Jan Behrens