Thanks Thomas et al,
your advice and pointer have been of great help. to minimise my code change I've decide to use the RETURNS TABLE option, but I now have the following problem.
My procedure is as follows:
CREATE or REPLACE FUNCTION myproc(var1 CHAR(9), var2 (CHAR4)) RETURNS TABLE (result INTEGER, reply TEXT)) AS $$
DECLARE
replyx TEXT;
result INTEGER;
BEGIN
replyx := 'HELLO WORLD';
result := 150
RETURN QUERY SELECT result, replyx;
END;
$$ LANGUAGE plpgsql;
I call it from ECPG (EMBEDDED SQL IN C) as follows:
EXEC SQL BEGIN DECLARE SECTION;
int iResult
char acReply[1000];
EXEC SQL END DECLARE SECTION;
EXEC SQL SELECT myproc ('abcdefghi', 'test') INTO :iResult, :acReply;
The error I get when I run the code is:
SQLSTATE: 42804
ERROR MESSAGE: invalid input syntax for type int: "(150,"HELLO
What am I doing wrong?
Many Thanks
Atif
On 8 July 2010 12:53, Thomas Kellerer
<spam_eater@gmx.net> wrote:
Thomas Kellerer, 08.07.2010 13:43: Atif Jung, 08.07.2010 11:51:
Hi,
I'm having difficulty working out the correct syntax to return more than
one value from a stored procedure. I wish to return an INTGER and a
string
CREATE or REPLACE FUNCTION testproc(val1 (CHAR9), val2 CHAR(4)) RETURNS
INTEGER, CHAR(640) AS $$
The above is incorrect but what is the correct syntax?
Thanks
Atif
If you don't need the power of pl/pgSQL to calculate your result, a
simple SQL function should work:
CREATE OR REPLACE FUNCTION testproc(val1 char(9), val2 char(4))
RETURNS TABLE(id integer, some_value text)
AS
$$
SELECT 42, 'your value'::text;
$$
LANGUAGE sql;
I did hit "Send" too quickly...
If you do need calculations in there (and thus the power of PL/pgSQL), you can do that as well:
CREATE OR REPLACE FUNCTION testproc(val1 char(9), val2 char(4))
RETURNS TABLE(id integer, some_value text)
AS
$$
DECLARE
return_int integer;
return_text text;
BEGIN
return_int := 21 * 2;
return_text := 'Your input value: ' || val1;
RETURN QUERY SELECT return_int, return_text;
END
$$
LANGUAGE plpgsql;
Both can be used like this: select * from testproc('x', 'y')