postgres stored procedure problem - postgresql

Question

Ich have a problem in postgres function:
CREATE OR REPLACE FUNCTION getVar(id bigint)
RETURNS TABLE (repoid bigint, suf VARCHAR, nam VARCHAR)
AS $$
declare rec record;
BEGIN
FOR rec IN
(WITH RECURSIVE children(repoobjectid,variant_of_object_fk, suffix, variantname) AS (
SELECT repoobjectid, variant_of_object_fk, '' as suffix,variantname
FROM b2m.repoobject_tab
WHERE repoobjectid = id
UNION ALL
SELECT repo.repoobjectid, repo.variant_of_object_fk, suffix || '..' , repo.variantname
FROM b2m.repoobject_tab repo, children
WHERE children.repoobjectid = repo.variant_of_object_fk)
SELECT repoobjectid,suffix,variantname FROM children)
LOOP
RETURN next;
END LOOP;
RETURN;
END;
It can be compiled, but if y try to call it
select * from getVar(18)
I got 8 empty rows with 3 columns.
If i execute the following part of procedure with hard-coded id parameter:
WITH RECURSIVE children(repoobjectid,variant_of_object_fk, suffix, variantname) AS (
SELECT repoobjectid, variant_of_object_fk, '' as suffix,variantname
FROM b2m.repoobject_tab
WHERE repoobjectid = 18
UNION ALL
SELECT repo.repoobjectid, repo.variant_of_object_fk, suffix || '..' , repo.variantname
FROM b2m.repoobject_tab repo, children
WHERE children.repoobjectid = repo.variant_of_object_fk)
SELECT repoobjectid,suffix,variantname FROM children
I got exactly, what i need 8 rows with data:
repoobjectid suffix variantname
18
19 .. for IPhone
22 .. for Nokia
23 .... OS 1.0
and so on.
What is going wrong ? Please help.
Thanx in advance

Answer 1

I think if you're doing "return table", you need to assign to the "columns" of the table before doing "return next". So something like:
repoid := rec.repoid;
suf := rec.suf;
nam := rec.nam;
just before your "RETURN NEXT". Since you're not assigning these, they're being returned as null.

Answer 2

Here is a sample code of a funcion that returns rows from a table, I think it might help.
First, a sample table with sample data:
CREATE TABLE sample_table (id smallint, description varchar, primary key (id));
INSERT INTO sample_table (id, description) VALUES (1, 'AAAA');
INSERT INTO sample_table (id, description) VALUES (2, 'BBBB');
INSERT INTO sample_table (id, description) VALUES (3, 'CCCC');
INSERT INTO sample_table (id, description) VALUES (4, 'DDDD');
INSERT INTO sample_table (id, description) VALUES (5, 'EEEE');
Then, a return type that describes the fields of the rows returned:
CREATE TYPE return_type AS
(id smallint,
description varchar);
ALTER TYPE return_type OWNER TO postgres;
Then the function itself:
CREATE OR REPLACE FUNCTION report(p_id integer)
RETURNS SETOF return_type AS
$BODY$
DECLARE
retorno return_type%ROWTYPE;
BEGIN
FOR RETORNO IN SELECT * FROM sample_table WHERE id = p_id LOOP
RETURN NEXT RETORNO;
END LOOP;
RETURN;
END
$BODY$
LANGUAGE 'plpgsql' VOLATILE
COST 100
ROWS 1000;
ALTER FUNCTION report(p_id integer) OWNER TO postgres;
And here goes the function call:
SELECT * FROM report(1);