SQL & PL/SQL

I was trying to find a fast way to optimize a script that insert a large amount of records into a table. The initial script was like

insert into table_xxxx
select a.camp1, a.camp2, a.camp3 a.camp4, b.camp1, b.camp2, b.camp3
from table_a a, table_b b
where a.camp0 = b.camp0

The commit sentence was at the end of the insert script; so i came up with this solution

Declare
TYPE cur_CURSOR IS REF CURSOR ;
TYPE Tab_Hist IS TABLE OF table_xxxx%ROWTYPE INDEX BY BINARY_INTEGER;
g_tHist Tab_Hist;

CURSOR c_Base IS
select a.camp1, a.camp2, a.camp3 a.camp4, b.camp1, b.camp2, b.camp3
from table_a a, table_b b
where a.camp0 = b.camp0;

BEGIN
OPEN c_base;
LOOP
FETCH c_base BULK COLLECT INTO g_tHist LIMIT 1000;
EXIT WHEN g_tHist.COUNT = 0;
BEGIN
FORALL i IN g_tHist.FIRST .. g_tHist.COUNT SAVE EXCEPTIONS
INSERT INTO prov_cobr_dud VALUES g_tHist(i);
COMMIT;
EXCEPTION
WHEN NO_DATA_FOUND THEN
NULL;
END;
g_tHist.DELETE;
EXIT WHEN c_base%NOTFOUND;
END LOOP;
CLOSE c_base;
----
commit;
----

END;

If anyone could tell me another way to do the same thing i'll apreciate it a lot; i'm keen on learn more efficient ways to optimize scripts.

PD: The initial insert was inserting the table with 120,000 records (more or less)