FORALL bulk delete statement it requires lot of time

Question

Hi,
when I execute FORALL bulk delete statement it requires lot of time even for 10 rows deletion.so how to avoid this problem? I checked SGA_TARGET and PGA_TARGET cureent_size and max_size are same for both. Is their is memory problem?
thanks in advance
Vaibhav

Answer 1

Please look at this:
http://psoug.org/definition/LIMIT.htm
If you add the LIMIT 100 part to your query your bulk collect / forall construction will process 100 records per part.
This will avoid filling up undo tablespaces / PGA memory.
There's also a discussion wether to use LIMIT 100 or LIMIT 1000 (or greater).
It depends on what your system can handle, if it's a busy production statement I'd stick to a value of 100.
Reason for this, is that your query will not take a huge amount of PGA.
DECLARE
TYPE t_id_tab IS TABLE OF test.c1%TYPE;
l_id_tab t_id_tab := t_id_tab();
Begin
dbms_output.put_line(DBMS_UTILITY.get_time);
select c1 bulk collect into l_id_tab
from TEST where c1<=150000
LIMIT 100;
FORALL i IN l_id_tab.first .. l_id_tab.last
delete from TEST where c1= l_id_tab(i);
commit;
dbms_output.put_line(DBMS_UTILITY.get_time);
End;we use this construction to load data into our warehouse and bumped into the LIMIT 100 part while testing.
It did speed up things significantly.
Another (risky!) trick is to do:
-- exclude table from archivelogging
ALTER TABLE TEST NOLOGGING;
-- execute plsql block
<your code comes here>
-- enable logging on the table
ALTER TABLE TEST LOGGING;The risky part is that you get inconsistent archivelogging this way because your deletes arent logged.
When you do a restore from backup and want to roll forward until the latest change in your archive logs, you'll miss these deletes and your table will not be consistent with the rest of your database.
In our warehouse situation we dont mind, data gets refreshed every night.
Hope this helps!
Robin