Issue with BULK COLLECT of million rows - out of process memory - oracle

Question

I have below code in my pl/sql procedure, which is trying to collect 17 millions rows, and it fails with error ORA-04030: out of process memory when trying to allocate 16328 bytes (koh-kghu call ,pmuccst: adt/record)
TYPE rc_test IS RECORD(
ROWID VARCHAR2(200),
contact_id VARCHAR2(200),
last_name VARCHAR2(200),
first_name VARCHAR2(200),
phone VARCHAR2(200),
email VARCHAR2(200),
birth_day date,
address_id VARCHAR2(200),
seq NUMBER,
NEWID NUMBER);
TYPE rctype
IS TABLE OF RC_TEST;
rcrecords RCTYPE;
BEGIN
SELECT ROWID,
contact_id,
last_name,
first_name,
phone,
email,
birth_day,
address_id,
seq,
NEWID
bulk collect INTO rcrecords
FROM HR.TMP_TBL_SEQ a
order by last_name, first_name, seq;
FOR i IN 1..rcrecords.count LOOP
-- <<>>
END LOOP;
It runs fine with sample data of hundreds or thousand of rows but doesn't work with millions of rows.
I have read troubleshooting for this error but most of them revolve around getting this done through DBA who will perform some tunings or will advise us to increase RAM that might resolve the issue but due to certain limitations at first place I really want to change the logic of the code itself.
Is there a better way to improve the logic in the above type of code to get rid of the out of memory error?
In place of table record used in my code can anyone provide me alternate code?
Many Thanks.

Answer 1

You can use BULK COLLECT with LIMIT clause (http://www.oracle.com/technetwork/issue-archive/2008/08-mar/o28plsql-095155.html). In your case it will look similar to :
CURSOR cur1 IS
SELECT ROWID, contact_id, last_name, first_name,
phone, email, birth_day, address_id, seq,
NEWID FROM HR.TMP_TBL_SEQ a order by last_name, first_name, seq;
TYPE T_TEST IS TABLE OF cur1 %ROWTYPE INDEX BY PLS_INTEGER;;
test_data T_TEST;
BEGIN
OPEN cur1;
LOOP
FETCH cur1
BULK COLLECT INTO test_data LIMIT 10000;
FOR indx IN 1 .. test_data.COUNT
LOOP
NULL;
--- process data
END LOOP;
EXIT WHEN test_data.COUNT < 10000;
END LOOP;
CLOSE cur1;

Answer 2

Avoid this sort of "unlimited" use of BULK COLLECT by using the LIMIT clause.
Move the SELECT statement into an explicit cursor declaration and then use a simple loop to fetch many but not all the rows from the table with each execution of the loop body using the optional LIMIT clause.
For an examples and explanations see this oracle tech-issue, see http://www.oracle.com/technetwork/issue-archive/2008/08-mar/o28plsql-095155.html

Answer 3

Replace the BULK COLLECT with a cursor for-loop:
begin
for rcrecords in
(
SELECT ROWID,
contact_id,
last_name,
first_name,
phone,
email,
birth_day,
address_id,
seq,
NEWID
bulk collect INTO rcrecords
FROM HR.TMP_TBL_SEQ a
order by last_name, first_name, seq
) loop
--Process data.
null;
end loop;
end;
/
Oracle will automatically take care of bulk collect, limit, and declaring the the data types.