From: | Stephan Szabo <sszabo(at)megazone(dot)bigpanda(dot)com> |
---|---|
To: | Christian Schoenebeck <cschoene(at)stud(dot)hs-heilbronn(dot)de> |
Cc: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: CASE statement and SETOF values |
Date: | 2006-07-24 22:01:35 |
Message-ID: | 20060724145053.O65749@megazone.bigpanda.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Mon, 24 Jul 2006, Christian Schoenebeck wrote:
> Consider the following server side function:
>
> CREATE FUNCTION my_function(int4) RETURNS SETOF int8 AS
> $BODY$
> SELECT
> CASE WHEN (some_condition)
> THEN (
> SELECT ... -- arbitrary select (returning row(s) of int8 values)
> )
> ELSE (
> SELECT ... -- arbitrary select (returning row(s) of int8 values)
> )
> END
> $BODY$
> LANGUAGE 'sql' VOLATILE;
>
> This function works fine if one of the two inner SELECT statements returns
> exactly one result (one row), but fails whenever one of them returns more
> than one result / rows.
>
> What is the reason? I mean the function is declared as returning "SETOF int8",
> so why does it expect a scalar?
The above basically looks like:
CASE WHEN <search condition> THEN <value expression> ELSE
<value expression> END.
In SQL92 at least, the form of <value expression> which looks like (SELECT
...) is <scalar subquery> which is limited to 1 column and 1 row. The
other subquery forms don't look legal in that position unless they changed
that in a later version of the spec.
From | Date | Subject | |
---|---|---|---|
Next Message | Merlin Moncure | 2006-07-25 01:20:12 | Re: Beginner's questions about creating a custom data type in PostgreSQL... |
Previous Message | Roderick A. Anderson | 2006-07-24 21:36:11 | Does a GRANT on a table cascade/implied to its SEQUENCES |