Very very long execution time

Question

hi,
working with 10g XE
new_header table has 9749 rows and consist of two column : itemsets and codes like that
AST:GLU:KREA:Na:IBIL, 2380. codes starts from 1 to 9749
when I run belows pl/sql codes, it takes very long time in fact it can not be complete without error
declare
cnt1 number(15) ;
cnt2 number(15) ;
itemset1 varchar(800) ;
itemset2 varchar2(800) ;
begin
select max(codes) into cnt1 from new_header ;
cnt2 := cnt1 ;
for f in 1..cnt1 loop --cnt1 loop
for l in 1..cnt2 loop --cnt2 loop
itemset1 := 'AST:GLU:KREA:Na:TBIL:IBIL' ; ---given for exeample
--select itemsets into itemset1 from new_header where codes = f ;
insert into k1
select REGEXP_SUBSTR (itemset1,'[^:]+', 1, level)
from dual
connect by level <= length(regexp_replace(itemset1 ,'[^:]*'))+1 ;
select itemsets into itemset2 from new_header where codes = l ;
insert into k2
select REGEXP_SUBSTR (itemset2,'[^:]+', 1, level)
from dual
connect by level <= length(regexp_replace(itemset2 ,'[^:]*'))+1 ;*/
check_it ;
commit ;
end loop ;
end loop ;
dbms_output.put_line(myres.count) ;
end ;
I wanto to get each part of itemset1 seperated by : k1 global temporary table has one column varchar2 will be used in check_it procudere to compare with a table values. please give me some advice urgent..what can i do. what can I use instead of regexp_substr or any idea..thanks..

Answer 1

The regexp_substr isn't the problem.
The problem is your table has 9749 rows and you're doing this:
for f in 1..cnt1 loop --cnt1 loop
for l in 1..cnt2 loop --cnt2 loop
end loop ;
end loop ;i.e. you are trying to execute the code in the loops 9749^2 which is roughly 95 million times. And you wonder why its taking a long time.
I havn't even tried to work out what you are trying to achieve in the rest of the code - maybe you could elaborate.
I suggest you rethink your approach to the problem though. Oh, and no, it's not urgent.