In my previous article, I explained PL/SQL records. Now we shall combine RECORDs with TABLEs to achieve effective results in a simple way. Let us consider the following example.

declare
    type t_empRec is record
    (
        ename emp.ename%type,
        sal emp.sal%type,
        deptno emp.deptno%type
    );
    type t_emptbl is table of t_emprec;
    v_emptbl    t_emptbl;
begin
    select ename,sal,deptno BULK COLLECT into v_emptbl from emp;
    for i in v_emptbl.first .. v_emptbl.last
    loop
        dbms_output.put_line(v_emptbl(i).ename || ',' ||
         v_emptbl(i).sal || ',' || v_emptbl(i).deptno);
    end loop;
end;

The above program retrieves 'ename', 'sal' and 'deptno' columns from 'emp' table and displays all of those details. In my previous article, I displayed the same but used %ROWTYPE. Now in this program, I am combining the definitions of RECORD and TABLE to store only the data we need (but not the entire row). The most important statement in the above program is the following:

type t_emptbl is table of t_emprec;

That statement defines a PL/SQL table named 't_emptbl'. But the content (rows) within that table should match with the structure defined in the following declaration:

type t_empRec is record
(
    ename emp.ename%type,
    sal emp.sal%type,
    deptno emp.deptno%type
);

So, indirectly 't_emptbl' can have any number of rows with only the fields 'ename', 'sal' and 'deptno'. This is a wonderful technique to define PL/SQL tables with our own fields. The rest of the program is just similar to the example I gave in my previous article.