Faster procedure that can be use

Question

pls guy i need to knw hw i ll b able to get a better procedure to run for shorter time.
i am trying extract data from two tables one in a remote location and another one in the present databse i m working on.
i want to use the extracted dat to update a column in the table.can u help to tune it better.i need to update a table of records of abt 70 million records.this is the procedure
CREATE OR REPLACE procedure AC_DATA.post_emp IS
begin
for i in (select distinct emp_id, user_id, user_emp_id
from emp, cust@miami
where emp_id = user_id)loop
begin
--update emp_log
update /*+ PARALLEL(emp_log,20) */ emp_log
set emp_id = i.user_emp_id
where emp_id = i.user_id;
commit;
end;
end loop;
end;
/

Answer 1

olayide wrote:
@ justin,tanx i am trying to run a procedure that will update the value of a particular column by comparing between two columns.I understand that you are doing something like the correlated update I posted. Again, since you haven't posted your table definitions, your data, or your expected output, it's rather hard for us to know exactly what correlated update you really want.
If you really want to do row-by-row processing (also known as "slow by slow" processing), it will take substantially longer to make the changes. It's also completely pointless to ask for parallel processing to be used for a single-row update. If you're not concerned about performance, loops can be very easy to code. But it sounded like you were concerned about performance.
Committing in the loop will, similarly, add an order of magnitude to the time required to process the data. And you don't appear to be doing anything that would make your code restartable in order to gain any sort of benefit from interim commits.
Justin