Pipeline table function - oracle

Question

I would like to understand why pipelined function is not returning any results
Any ideas what i am doing wrong here.
Giving credit where due to the Original Author
CREATE OR REPLACE PACKAGE MANAGE_SPACE AS
--Author Tanmay
g_tblspce_threshold number := 80;
TYPE tblespaces_record IS RECORD(
tablespace_name VARCHAR2(30),
percentage_used NUMBER
);
TYPE tblespaces_table IS TABLE OF tblespaces_record;
function list_tblspcs_excd_thresld
return tblespaces_table PIPELINED ;
END MANAGE_SPACE;
Package body
CREATE OR REPLACE PACKAGE BODY MANAGE_SPACE
AS
--Author Tanmay
FUNCTION list_tblspcs_excd_thresld
RETURN tblespaces_table PIPELINED
AS
tblspaces tblespaces_record;
BEGIN
for x in (SELECT a.tablespace_name tablespace_name,
ROUND (((c.BYTES - NVL (b.BYTES, 0)) / c.BYTES) * 100) percentage_used
into tblspaces
FROM dba_tablespaces a,
(SELECT tablespace_name,
SUM (BYTES) BYTES
FROM dba_free_space
GROUP BY tablespace_name
) b,
(SELECT COUNT (1) DATAFILES,
SUM (BYTES) BYTES,
tablespace_name
FROM dba_data_files
GROUP BY tablespace_name
) c
WHERE b.tablespace_name(+) = a.tablespace_name
AND c.tablespace_name(+) = a.tablespace_name
AND ROUND (((c.BYTES - NVL (b.BYTES, 0)) / c.BYTES) * 100) > g_tblspce_threshold
ORDER BY NVL (((c.BYTES - NVL (b.BYTES, 0)) / c.BYTES), 0) DESC)
loop
PIPE ROW (tblspaces);
end loop;
return;
END list_tblspcs_excd_thresld;
END MANAGE_SPACE;
executing this package does not return any rows
SQL> select * from table(MANAGE_SPACE.list_tblspcs_excd_thresld())
2
SQL> /
TABLESPACE_NAME PERCENTAGE_USED
------------------------------ ---------------
What am i doing wrong

Answer 1

You need to populate an array. The easiest way to do this is to use the BULK COLLECT syntax. Then loop round the array and pipe out rows.
Here is my revised version of your package.
CREATE OR REPLACE PACKAGE BODY MANAGE_SPACE
AS
FUNCTION list_tblspcs_excd_thresld
RETURN tblespaces_table PIPELINED
AS
-- collection type not record type
tblspaces tblespaces_table;
BEGIN
select *
-- populate the collection
bulk collect into tblspaces
from
(SELECT a.tablespace_name tablespace_name,
ROUND (((c.BYTES - NVL (b.BYTES, 0)) / c.BYTES) * 100) percentage_used
FROM dba_tablespaces a,
(SELECT tablespace_name,
SUM (BYTES) BYTES
FROM dba_free_space
GROUP BY tablespace_name
) b,
(SELECT COUNT (1) DATAFILES,
SUM (BYTES) BYTES,
tablespace_name
FROM dba_data_files
GROUP BY tablespace_name
) c
WHERE b.tablespace_name(+) = a.tablespace_name
AND c.tablespace_name(+) = a.tablespace_name
AND ROUND (((c.BYTES - NVL (b.BYTES, 0)) / c.BYTES) * 100) > g_tblspce_threshold
ORDER BY NVL (((c.BYTES - NVL (b.BYTES, 0)) / c.BYTES), 0) DESC);
-- loop round the collection
for i in 1..tblspaces.count()
loop
PIPE ROW (tblspaces(i));
end loop;
return;
END list_tblspcs_excd_thresld;
END MANAGE_SPACE;
Here is the output:
SQL> select * from table(MANAGE_SPACE.list_tblspcs_excd_thresld())
2 /
TABLESPACE_NAME PERCENTAGE_USED
------------------------------ ---------------
SYSTEM 100
SYSAUX 91
APEX_2614203650434107 90
EXAMPLE 87
USERS 82
SQL>
#a_horse_with_no_name makes a good point regarding memory consumption. Collections are held in the session's memory space, the PGA. In this particular case it is unlikely that a database would have enough tablespaces to blow the PGA limit. But for other queries which might have much larger result sets there is the BULK COLLECT ... LIMIT syntax, which allows us to fetch manageable chunks of the result set into our collection. Find out more.