I want to creat a primary for a table,for example,there is a table named "test" which include field a1,a2,a3,...,I want to make field a1 and a2 to become a primary key,but there is a question,I make primary key failed,because combination of field a1 and a2 raise some same records.I want to know how to erase these same records? The same records are about 5000 records,I can't delete them manually,I want to delete them automatically,for example,if there is five same records,I delete four,and reserve one. How to realize this function? I am a newbie for Oracle,please give detail steps.
Thanks!

Dear friend,
You can try this to delete all repeated records

delete test t1
where (select count(*) from test t2 where t1.a1=t2.a1 and t1.a2=t2.a2) >1;

and if you want to remain one record from the repeated records you can try this
tigger :

declare
cursor c1 is
select a1,a2,count(*) from test group by a1,a2;
 begin
 for c1_r in c1 loop
   if c1_r.co >1 then
    for i in 1..(c1_r.co-1) loop
      delete test
       where test.a1=c1_co.a1 and test.a2=c1_co.a2 and rownum=1;
    end loop;
end loop;
forms_ddl('commit');
end;


Waheed rashwan

Simply use,
delete from test where rowid not in (select min(rowid) from test group by a1, a2);

It is the most fastest and efficient way to delete all the duplicate records.
Sujit