Duplicate records in a collection

Question

Hi Experts,
Just now I've seen a thread related to finding duplicate records in a collection. I understand that it is not advisable to sort/filter data in a collection.
(https://forums.oracle.com/thread/2584168)
Just for curiosity I tried to display duplicate records in a collection. Please Please .. this is just for practice purpose only. Below is the rough code which I wrote.
I'm aware of one way - can be handled effectively by passing data into a global temporary table and display the duplicate/unique records.
Can you please let me know if there is any other efficient wayto do this.
declare
  type emp_rec is record ( ename varchar2(40), empno number);
  l_emp_rec emp_rec; 
  type emp_tab is table of l_emp_rec%type index by binary_integer;
  l_emp_tab emp_tab;
  l_dup_tab emp_tab;
  l_cnt number;
  n number :=1;
begin
-- Assigning values to Associative array
  l_emp_tab(1).ename := 'suri';
  l_emp_tab(1).empno := 1;
  l_emp_tab(2).ename := 'surya';
  l_emp_tab(2).empno := 2;
  l_emp_tab(3).ename := 'suri';
  l_emp_tab(3).empno := 1;
-- Comparing collection for duplicate records
for i in l_emp_tab.first..l_emp_tab.last
loop
    l_cnt :=0;  
for j in l_emp_tab.first..l_emp_tab.last 
    loop      
       if l_emp_tab(i).empno  =  l_emp_tab(j).empno and l_emp_tab(i).ename  =  l_emp_tab(j).ename then
           l_cnt := l_cnt+1;          
               if l_cnt >=2 then
                  l_dup_tab(n):= l_emp_tab(i);
               end if;
       end if;                   
    end loop;  
end loop;
-- Displaying duplicate records
for i in l_dup_tab.first..l_dup_tab.last
loop
   dbms_output.put_line(l_dup_tab(i).ename||'  '||l_dup_tab(i).empno);
end loop;
end;
Cheers,
Suri

Answer 1

Dunno if this is either easier or more efficient but it is different.  The biggest disadvantage to this technique is that you have extraneous database objects (a table) to keep track of.  The advantage is that you can use SQL to perform the difference checks easily.
Create 2 global temporary tables with the structure you need, load them, and use set operators (UNION [ALL], INTERSECT, MINUS) to find the differences.  Or, create 1 GTT with an extra column identifying the set and use the extra column to identify the set records you need.