pg.ddx.io pgsql-sql@postgresql.org mailing list archive
help / color / mirror / Atom feedSTRICT function returning a composite type
7+ messages / 3 participants
[nested] [flat]
* STRICT function returning a composite type
@ 2003-11-13 16:08 Alexander M. Pravking <fduch@antar.bryansk.ru>
0 siblings, 1 reply; 7+ messages in thread
From: Alexander M. Pravking @ 2003-11-13 16:08 UTC (permalink / raw)
To: pgsql-sql
I noted that such a function returns an empty rowset if a NULL value is
passed as an argument. Is it a bug or feature? I wish it was a feature,
because I probably want to use this behavour.
Here's an example:
CREATE TYPE ts_bounds AS (
sdate timestamptz,
edate timestamptz
);
CREATE OR REPLACE FUNCTION overlap_bounds(timestamptz, timestamptz, timestamptz, timestamptz)
RETURNS ts_bounds AS '
DECLARE
sdate1 ALIAS FOR $1;
edate1 ALIAS FOR $2;
sdate2 ALIAS FOR $3;
edate2 ALIAS FOR $4;
res ts_bounds%rowtype;
BEGIN
res.sdate := CASE WHEN sdate1 > sdate2 THEN sdate1 ELSE sdate2 END;
res.edate := CASE WHEN edate1 < edate2 THEN edate1 ELSE edate2 END;
IF res.sdate > res.edate THEN
res.sdate := NULL;
res.edate := NULL;
END IF;
RETURN res;
END' LANGUAGE 'plPgSQL' STRICT;
fduch=# SELECT * from overlap_bounds('-infinity', 'today', 'yesterday', 'infinity');
sdate | edate
------------------------+------------------------
2003-11-12 00:00:00+03 | 2003-11-13 00:00:00+03
(1 row)
fduch=# SELECT * from overlap_bounds('-infinity', 'today', 'yesterday', null);
sdate | edate
-------+-------
(0 rows)
What I want is to get no rows if given intervals don't overlap instead of:
fduch=# SELECT * from overlap_bounds('-infinity', 'yesterday', 'today', 'infinity');
sdate | edate
-------+-------
|
(1 row)
Is it possible without returning SETOF ts_bounds?
fduch=# SELECT version();
version
---------------------------------------------------------------------
PostgreSQL 7.3.4 on i386-portbld-freebsd4.8, compiled by GCC 2.95.4
--
Fduch M. Pravking
^ permalink raw reply [nested|flat] 7+ messages in thread
* Re: STRICT function returning a composite type
@ 2003-11-13 17:14 Richard Huxton <dev@archonet.com>
parent: Alexander M. Pravking <fduch@antar.bryansk.ru>
0 siblings, 1 reply; 7+ messages in thread
From: Richard Huxton @ 2003-11-13 17:14 UTC (permalink / raw)
To: Alexander M. Pravking <fduch@antar.bryansk.ru>; pgsql-sql
On Thursday 13 November 2003 16:08, Alexander M. Pravking wrote:
> I noted that such a function returns an empty rowset if a NULL value is
> passed as an argument. Is it a bug or feature? I wish it was a feature,
> because I probably want to use this behavour.
^ permalink raw reply [nested|flat] 7+ messages in thread
* Re: STRICT function returning a composite type
@ 2003-11-13 17:19 Alexander M. Pravking <fduch@antar.bryansk.ru>
parent: Richard Huxton <dev@archonet.com>
0 siblings, 1 reply; 7+ messages in thread
From: Alexander M. Pravking @ 2003-11-13 17:19 UTC (permalink / raw)
To: Richard Huxton <dev@archonet.com>; +Cc: pgsql-sql
On Thu, Nov 13, 2003 at 05:14:27PM +0000, Richard Huxton wrote:
> RETURNS NULL ON NULL INPUT or STRICT indicates that the function always
> returns NULL whenever any of its arguments are NULL. If this parameter is
> specified, the function is not executed when there are NULL arguments;
> instead a NULL result is assumed automatically.
Does "NULL result" mean an empty rowset if the function returns a record?
--
Fduch M. Pravking
^ permalink raw reply [nested|flat] 7+ messages in thread
* Re: STRICT function returning a composite type
@ 2003-11-13 17:27 Tom Lane <tgl@sss.pgh.pa.us>
parent: Alexander M. Pravking <fduch@antar.bryansk.ru>
0 siblings, 1 reply; 7+ messages in thread
From: Tom Lane @ 2003-11-13 17:27 UTC (permalink / raw)
To: Alexander M. Pravking <fduch@antar.bryansk.ru>; +Cc: Richard Huxton <dev@archonet.com>; pgsql-sql
"Alexander M. Pravking" <fduch@antar.bryansk.ru> writes:
> On Thu, Nov 13, 2003 at 05:14:27PM +0000, Richard Huxton wrote:
>> RETURNS NULL ON NULL INPUT or STRICT indicates that the function always
>> returns NULL whenever any of its arguments are NULL.
> Does "NULL result" mean an empty rowset if the function returns a record?
No, it means a null record. "Empty rowset" would apply to a function
declared to return SETOF something. (I believe that is how we
interpret the concept of strictness for functions returning sets.)
regards, tom lane
^ permalink raw reply [nested|flat] 7+ messages in thread
* Re: STRICT function returning a composite type
@ 2003-11-13 17:32 Alexander M. Pravking <fduch@antar.bryansk.ru>
parent: Tom Lane <tgl@sss.pgh.pa.us>
0 siblings, 1 reply; 7+ messages in thread
From: Alexander M. Pravking @ 2003-11-13 17:32 UTC (permalink / raw)
To: Tom Lane <tgl@sss.pgh.pa.us>; +Cc: Richard Huxton <dev@archonet.com>; pgsql-sql
On Thu, Nov 13, 2003 at 12:27:58PM -0500, Tom Lane wrote:
> "Alexander M. Pravking" <fduch@antar.bryansk.ru> writes:
> > Does "NULL result" mean an empty rowset if the function returns a record?
>
> No, it means a null record. "Empty rowset" would apply to a function
> declared to return SETOF something. (I believe that is how we
> interpret the concept of strictness for functions returning sets.)
Very well then... Can I return a null record from such function
explicitly? Sorry, I could't find it anywhere in docs or examples.
--
Fduch M. Pravking
^ permalink raw reply [nested|flat] 7+ messages in thread
* Re: STRICT function returning a composite type
@ 2003-11-13 17:35 Tom Lane <tgl@sss.pgh.pa.us>
parent: Alexander M. Pravking <fduch@antar.bryansk.ru>
0 siblings, 1 reply; 7+ messages in thread
From: Tom Lane @ 2003-11-13 17:35 UTC (permalink / raw)
To: Alexander M. Pravking <fduch@antar.bryansk.ru>; +Cc: Richard Huxton <dev@archonet.com>; pgsql-sql
"Alexander M. Pravking" <fduch@antar.bryansk.ru> writes:
> Very well then... Can I return a null record from such function
> explicitly? Sorry, I could't find it anywhere in docs or examples.
Not sure. Seems like you should be able to, but I've never tried it.
regards, tom lane
^ permalink raw reply [nested|flat] 7+ messages in thread
* Re: STRICT function returning a composite type
@ 2003-11-13 17:42 Alexander M. Pravking <fduch@antar.bryansk.ru>
parent: Tom Lane <tgl@sss.pgh.pa.us>
0 siblings, 0 replies; 7+ messages in thread
From: Alexander M. Pravking @ 2003-11-13 17:42 UTC (permalink / raw)
To: Tom Lane <tgl@sss.pgh.pa.us>; +Cc: Richard Huxton <dev@archonet.com>; pgsql-sql
On Thu, Nov 13, 2003 at 12:35:41PM -0500, Tom Lane wrote:
> "Alexander M. Pravking" <fduch@antar.bryansk.ru> writes:
> > Very well then... Can I return a null record from such function
> > explicitly? Sorry, I could't find it anywhere in docs or examples.
>
> Not sure. Seems like you should be able to, but I've never tried it.
Thanks for a quick response, guys. I'll try to find it myself and
will let you know if I did ;)
--
Fduch M. Pravking
^ permalink raw reply [nested|flat] 7+ messages in thread
end of thread, other threads:[~2003-11-13 17:42 UTC | newest]
Thread overview: 7+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2003-11-13 16:08 STRICT function returning a composite type Alexander M. Pravking <fduch@antar.bryansk.ru>
2003-11-13 17:14 ` Richard Huxton <dev@archonet.com>
2003-11-13 17:19 ` Alexander M. Pravking <fduch@antar.bryansk.ru>
2003-11-13 17:27 ` Tom Lane <tgl@sss.pgh.pa.us>
2003-11-13 17:32 ` Alexander M. Pravking <fduch@antar.bryansk.ru>
2003-11-13 17:35 ` Tom Lane <tgl@sss.pgh.pa.us>
2003-11-13 17:42 ` Alexander M. Pravking <fduch@antar.bryansk.ru>
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox