Re: pl/sql dup_val_on_index

insert into select where not exists
or
insert into select where not in

Don't *kludge* procedurally what you can do non-procedurally. (Or stick to flat files)

Also your 'solution' wouldn't help, as it would abort as soon as you get the first duplicate.
You would need to
for i in (select * from original_table ) loop begin
insert into .. values (i.c1, i.c2);
exception
when dup_val_on_index then null;
end;
end loop;
/

Needless to say you will end up with an unscalable app.

-- 
Sybrand Bakker
Senior Oracle DBA