Les05 - Set Operators

1 of 20
Objectives
1 Describe Set Operators

2 Use set operators to combine multiple queries into a single query

3 Control order of rows returned

2 of 20
Topics Covered
Set Operator Types and rules

Tables Used in the lesson  Employee and Job_history

UNION

UNION ALL

INTERSECT

MINUS

Matching SELECT statements

ORDER BY with set operators

3 of 20
Types - Union
UNION

UNION of all
the rows in A
With ALL the
rows in B
With NO RESULT is the Yellow- but duplicates not showing twice

DUPLICATES

4 of 20
EXAMPLE:
JOB_HISTORY Table keeps history of when an employee changes jobs

Records start date and end date of employees that switch jobs
Employees who are still in the same job will not appear here
The current job is shown in the EMPLOYEE table. Again, this shows history.

SELECT * FROM job_history;

EMPLOYEE_ID START_DATE END_DATE JOB_ID DEPARTMENT_ID
------------------ ------------------------- ------------------ ---------- ---
102 13-JAN-93 24-JUL-98 IT_PROG 60
101 21-SEP-89 27-OCT-93 AC_ACCOUNT 110
101 28-OCT-93 15-MAR-97 AC_MGR 110
201 17-FEB-96 19-DEC-99 MK_REP 20
114 24-MAR-98 31-DEC-99 ST_CLERK 50
122 01-JAN-99 31-DEC-99 ST_CLERK 50
200 17-SEP-87 17-JUN-93 AD_ASST 90
176 24-MAR-98 31-DEC-98 SA_REP 80
176 01-JAN-99 31-DEC-99 SA_MAN 80
200 01-JUL-94 31-DEC-98 AC_ACCOUNT 90

10 rows selected

EMPLOYEE table contains employee information.

This example, only show the employee_id, job_id, department_id  the common attributes.

SELECT employee_id, job_id, department_id

FROM employee;

EMPLOYEE_ID JOB_ID DEPARTMENT_ID

----------- ---------- -------------
100 AD_PRES 90
101 AD_VP 90
102 AD_VP 90
103 IT_PROG 60
104 IT_PROG 60
107 IT_PROG 60
124 ST_MAN 50
141 ST_CLERK 50
142 ST_CLERK 50
143 ST_CLERK 50
144 ST_CLERK 50
149 SA_MAN 80
174 SA_REP 80
176 SA_REP 80
178 SA_REP
200 AD_ASST 10
201 MK_MAN 20
202 MK_REP 20
205 AC_MGR 110
206 AC_ACCOUNT 110
20 rows selected.

5 of 20
Result of the UNION of both tables using just
employee_id and job_id
SELECT employee_id, job_id
FROM employee
UNION
SELECT employee_id, job_id
FROM job_history
ORDER BY job_id;  added order by for readability

EMPLOYEE_ID JOB_ID
---------------------- ----------
100 AD_PRES
101 AC_ACCOUNT
101 AC_MGR
101 AD_VP
102 AD_VP
102 IT_PROG
103 IT_PROG
104 IT_PROG
107 IT_PROG
114 ST_CLERK
122 ST_CLERK
124 ST_MAN
141 ST_CLERK
142 ST_CLERK
143 ST_CLERK
144 ST_CLERK
149 SA_MAN
174 SA_REP
176 SA_MAN
176 SA_REP
178 SA_REP
200 AC_ACCOUNT
200 AD_ASST
201 MK_MAN
201 MK_REP
202 MK_REP
205 AC_MGR
206 AC_ACCOUNT

28 rows selected  Notice the 28 rows

Employee 20
Job_history 10
TOTAL 30 rows

Since only produced 28 rows, then 2 rows must be duplicates and not shown.

WHERE ARE THE DUPLICATES ?

6 of 20
Types – Union All
8-4

UNION of ALL
the rows in A
and B
including
duplicates
ENTER the code using the same 2 tables to see the result.

RESULT is all
the Yellow
and duplicates
showing twice

7 of 20
SELECT employee_id, job_id
FROM employee
UNION ALL
SELECT employee_id, job_id
FROM job_history
ORDER BY employee_id;

EMPLOYEE_ID JOB_ID
---------------------- ----------
100 AD_PRES
101 AD_VP
101 AC_ACCOUNT
101 AC_MGR
102 IT_PROG
102 AD_VP
103 IT_PROG
104 IT_PROG
107 IT_PROG
114 ST_CLERK
122 ST_CLERK
124 ST_MAN
141 ST_CLERK
142 ST_CLERK
143 ST_CLERK
144 ST_CLERK
149 SA_MAN
174 SA_REP
176 SA_REP Was a Sales Representative
176 SA_MAN Became a Sales Manager
176 SA_REP Went back to a Sales Rep
178 SA_REP
200 AD_ASST Looks like the same here
200 AD_ASST
200 AC_ACCOUNT
201 MK_REP
201 MK_MAN
202 MK_REP
205 AC_MGR
206 AC_ACCOUNT
30 rows selected

8 of 20
Change the code and add in DEPARTMENT_ID
SELECT employee_id, job_id, department_id
FROM employee
UNION
SELECT employee_id, job_id, department_id
FROM job_history
ORDER BY employee_id;

What was the result?

How many duplicates, if any?

Why?

ANSWER:

Look at the former duplicate 200 – notice different department_id

9 of 20
Types – Intersect
INTERSECT

10 of 20
 The rows in common to both tables only

A intersect B
same as
B intersect A

RESULT is the Yellow

Change the previous SQL to an INTERSECT

- finds the common rows (or duplicates)

SELECT employee_id, job_id, department_id

FROM employee
INTERSECT
SELECT employee_id, job_id, department_id
FROM job_history
ORDER BY employee_id;

EMPLOYEE_ID JOB_ID DEPARTMENT_ID

----------- ---------- -------------
176 SA_REP 80

11 of 20
Do the same but without department_id
SELECT employee_id, job_id
FROM employee
INTERSECT
SELECT employee_id, job_id
FROM job_history;

EMPLOYEE_ID JOB_ID
----------- ----------
176 SA_REP
200 AD_ASST

What did this tell you?

SELECT * FROM JOB_HISTORY;

EMPLOYEE_ID START_DAT END_DATE JOB_ID DEPARTMENT_ID

----------- --------- --------- ---------- -------------
102 13-JAN-93 24-JUL-98 IT_PROG 60
101 21-SEP-89 27-OCT-93 AC_ACCOUNT 110
101 28-OCT-93 15-MAR-97 AC_MGR 110
201 17-FEB-96 19-DEC-99 MK_REP 20
114 24-MAR-98 31-DEC-99 ST_CLERK 50
122 01-JAN-99 31-DEC-99 ST_CLERK 50
200 17-SEP-87 17-JUN-93 AD_ASST 90
176 24-MAR-98 31-DEC-98 SA_REP 80
176 01-JAN-99 31-DEC-99 SA_MAN 80
200 01-JUL-94 31-DEC-98 AC_ACCOUNT 90
10 rows selected.

12 of 20
TITLES and ORDER BY

SELECT employee_id as "Emp#", job_id as "Job Title"

FROM employee
UNION ALL
SELECT employee_id, job_id
FROM job_history
ORDER BY 1, 2

Emp# Job Title

---------- ----------
100 AD_PRES
101 AC_ACCOUNT
101 AC_MGR …. Etc for 30 rows

What if use 3 columns in table 1 and 2 in table 2?

SELECT employee_id as "Emp#", job_id as "Job Title", department_id

FROM employee
UNION ALL
SELECT employee_id, job_id
FROM job_history
ORDER BY 1, 2

Can’t make the comparison properly

SQL Error: ORA-01789: query block has incorrect number of result columns
01789. 00000 - "query block has incorrect number of result columns"

13 of 20
Types – Minus
8-4

MINUS

Rows in the
first query A
That are not in
second query
B
PRECEDENCE – equal – evaluated left to right

RESULT is Caution recommended. Use brackets with INTERSECT

the Yellow Alternate way of saying it is  Those rows that are unique to the first
query

SELECT employee_id, job_id

FROM employee
MINUS
SELECT employee_id, job_id
FROM job_history
ORDER BY 1, 2

Predict how many rows?

14 of 20
Table A or employees has 20-0  means all 20 rows with no duplicates
The intersect of A and B is 2 rows duplicated  result so far is 20 – 0 – 2 = 18

15 of 20
A bit more
Give a list of department_id, location_id, hire_date.

That requires 2 tables, EMPLOYEES and DEPARTMENT

Using a JOIN

SELECT E.department_id, location_id, hire_date

FROM employee E, department D
WHERE E.department_id = D.department_id

DEPARTMENT_ID LOCATION_ID HIRE_DATE

------------- ----------- ---------
10 1700 17-SEP-87
20 1800 17-FEB-96
20 1800 17-AUG-97
50 1500 16-NOV-99
50 1500 17-OCT-95
50 1500 29-JAN-97
50 1500 15-MAR-98
50 1500 09-JUL-98
60 1400 03-JAN-90
60 1400 21-MAY-91
60 1400 07-FEB-99
80 2500 29-JAN-00
80 2500 11-MAY-96
80 2500 24-MAR-98
90 1700 17-JUN-87
90 1700 21-SEP-89
90 1700 13-JAN-93
110 1700 07-JUN-94
110 1700 07-JUN-94

19 rows selected

16 of 20
SAME EXAMPLE but using UNION
Display department ID, location ID and hire date for all members

To use SET operators you need the same number of columns

Need 3 columns in employees

Need same 3 columns in departments

PROBLEM:

 Need hire_date from employees but it doesn't have a location_id in employees

 Need location_id from departments but it doesn't have a date to match with

17 of 20
SOLUTION

Because the expressions in the SELECT lists of the queries must match in number,
- use the dummy columns and the data type conversion functions to comply with this rule.

You must match the data type when columns do not exist in one or the other table
- use the TO_CHAR or any other conversion function to get the same data type

SELECT department_id, TO_NUMBER (null) as location, hire_date

FROM employee
UNION
SELECT department_id, location_id, TO_DATE (null) Note the location because
FROM department; TO_NUMBER (null) does
not make a good column
DEPARTMENT_ID LOCATION HIRE_DATE heading
------------- ---------- ---------
10 1700
10 17-SEP-87
20 1800
20 17-FEB-96
20 17-AUG-97
50 1500
50 17-OCT-95
50 29-JAN-97
50 15-MAR-98 From a diffeent example. Notice that the
50 09-JUL-98 blanks are nulls
50 16-NOV-99
60 1400
60 03-JAN-90
60 21-MAY-91
60 07-FEB-99
80 2500
80 11-MAY-96
80 24-MAR-98
80 29-JAN-00
90 1700
90 17-JUN-87
90 21-SEP-89
90 13-JAN-93
110 1700
110 07-JUN-94
190 1700
24-MAY-99

27 rows selected.

18 of 20
Matching SELECT statements
8-25

EXAMPLE 1:
Display all employees their job id and salary.

What are the problems?

Employees have several jobs and to display all the jobs requires a join to the job_history table
But … the job_history table does not have salary

SELECT employee_id, job_id, salary

FROM employee
UNION
Matching columns
SELECT employee_id, job_id, 0 If no salary will show 0
FROM job_history;

EMPLOYEE_ID JOB_ID SALARY

----------- ---------- ----------
100 AD_PRES 24000
101 AC_ACCOUNT 0
101 AC_MGR 0
101 AD_VP 17000
102 AD_VP 17000
102 IT_PROG 0  means, no record of salary of previous job
103 IT_PROG 9000
104 IT_PROG 6000
107 IT_PROG 4200
114 ST_CLERK 0  what does this mean?
122 ST_CLERK 0
124 ST_MAN 5800
141 ST_CLERK 3500
142 ST_CLERK 3100
143 ST_CLERK 2600
144 ST_CLERK 2500
149 SA_MAN 10500
174 SA_REP 11000
176 SA_MAN 0  means, no record of salary of previous job
176 SA_REP 0
176 SA_REP 8600
178 SA_REP 7000
200 AC_ACCOUNT 0
200 AD_ASST 0
200 AD_ASST 4400
201 MK_MAN 13000
201 MK_REP 0
202 MK_REP 6000
205 AC_MGR 12000
206 AC_ACCOUNT 8300

30 rows  20 + 10 - 0

19 of 20
Rules or Guidelines
8-5

- The expressions in the SELECT lists must match in number.

- If you select 3 columns in A, then must have 3 columns in B

- The data type of each column in the second query must match the data type
of its corresponding column in the first query.

- Parentheses can be used to alter the sequence of execution.

- ORDER BY clause can appear only at the very end of the statement.

Other

• Duplicate rows are automatically eliminated except in UNION ALL.

• Column names from the first query are the ones that appear in the result.

• The output is sorted in ascending order by default except in UNION ALL.

20 of 20