CREATE TABLE employee2

SELECT * FROM employee;

DECLARE

TYPE employee_tt IS TABLE OF employee%ROWTYPE;

group1 employee_tt;

group2 employee_tt;

l_count INTEGER;

t1 INTEGER;

PROCEDURE load1 (e IN OUT employee_tt)

IS

BEGIN

SELECT *

BULK COLLECT INTO e

FROM employee;

END;

PROCEDURE load2 (e IN OUT employee_tt)

IS

BEGIN

SELECT *

BULK COLLECT INTO e

FROM employee2;

END;

-- We can use = or !=, but not < or >.

t1 := DBMS_UTILITY.get_time;

load1 (group1);

load2 (group2);

FOR n IN 1 .. 1000

LOOP

IF group1 = group2

THEN

NULL;

END IF;

END LOOP;

DBMS_OUTPUT.put_line (DBMS_UTILITY.get_time - t1);

t1 := DBMS_UTILITY.get_time;

FOR n IN 1 .. 1000

LOOP

SELECT COUNT (*)

INTO l_count

FROM ((SELECT *

FROM employee

MINUS

SELECT *

FROM employee2)

UNION

(SELECT *

FROM employee2

MINUS

SELECT *

FROM employee));

IF l_count = 0

THEN

NULL;

END IF;

END LOOP;

DBMS_OUTPUT.put_line (DBMS_UTILITY.get_time - t1);

END;