Using the Set Operators

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Objectives

After completing this lesson, you should be able to do the

following:
• Describe set operators
• Use a set operator to combine multiple queries into a
single query
• Control the order of rows returned

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

In this lesson, you learn how to write queries by using set operators.
 Lesson Agenda

• Set Operators: Types and guidelines

• Tables used in this lesson
• UNION and UNION ALL operator
• INTERSECT operator
• MINUS operator
• Matching the SELECT statements
• Using the ORDER BY clause in set operations

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Set Operators
A B A B

UNION/UNION ALL

A B

INTERSECT

A B

MINUS

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

Set operators combine the results of two or more component queries into one result. Queries
containing set operators are called compound queries.

Operator Returns
UNION Rows from both queries after eliminating duplications
UNION ALL Rows from both queries, including all duplications
INTERSECT Rows that are common to both queries
MINUS Rows in the first query that are not present in the second query

All set operators have equal precedence. If a SQL statement contains multiple set operators,
the Oracle server evaluates them from left (top) to right (bottom) - if no parentheses explicitly
specify another order. You should use parentheses to specify the order of evaluation explicitly
in queries that use the INTERSECT operator with other set operators.
 Set Operator Rules

• The expressions in the SELECT lists must match in

number.
• The data type of each column in the subsequent 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.

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

• The expressions in the SELECT lists of the queries must match in number and data type.
Queries that use UNION, UNION ALL, INTERSECT, and MINUS operators must have the
same number and data type of columns in their SELECT list. The data type of the
columns in the SELECT list of the queries in the compound query may not be exactly the
same. The column in the second query must be in the same data type group (such as
numeric or character) as the corresponding column in the first query.
• Set operators can be used in subqueries.
• You should use parentheses to specify the order of evaluation in queries that use the
INTERSECT operator with other set operators. This ensures compliance with emerging
SQL standards that will give the INTERSECT operator greater precedence than the
other set operators.
 Oracle Server and Set Operators

• Duplicate rows are automatically eliminated except in

UNION ALL.
• Column names from the first query appear in the result.
• The output is sorted in ascending order by default except
in UNION ALL.

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

When a query uses set operators, the Oracle server eliminates duplicate rows automatically
except in the case of the UNION ALL operator. The column names in the output are decided
by the column list in the first SELECT statement. By default, the output is sorted in ascending
order of the first column of the SELECT clause.
The corresponding expressions in the SELECT lists of the component queries of a compound
query must match in number and data type. If component queries select character data, the
data type of the return values is determined as follows:
• If both queries select values of CHAR data type, of equal length, the returned values
have the CHAR data type of that length. If the queries select values of CHAR with different
lengths, the returned value is VARCHAR2 with the length of the larger CHAR value.
• If either or both of the queries select values of VARCHAR2 data type, the returned values
have the VARCHAR2 data type.
If component queries select numeric data, the data type of the return values is determined by
numeric precedence. If all queries select values of the NUMBER type, the returned values have
the NUMBER data type. In queries using set operators, the Oracle server does not perform
implicit conversion across data type groups. Therefore, if the corresponding expressions of
component queries resolve to both character data and numeric data, the Oracle server returns
an error.
 Tables Used in This Lesson

The tables used in this lesson are:

• EMPLOYEES: Provides details regarding all current
employees
• JOB_HISTORY: Records the details of the start date and
end date of the former job, and the job identification
number and department when an employee switches jobs

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

Two tables are used in this lesson: the EMPLOYEES table and the JOB_HISTORY table.
You are already familiar with the EMPLOYEES table that stores employee details such as a
unique identification number, email address, job identification (such as ST_CLERK, SA_REP,
and so on), salary, manager, and so on.
Some of the employees have been with the company for a long time and have switched to
different jobs. This is monitored using the JOB_HISTORY table. When an employee switches
jobs, the details of the start date and end date of the former job, the job_id (such as
ST_CLERK, SA_REP, and so on), and the department are recorded in the JOB_HISTORY
table.
The structure and data from the EMPLOYEES and JOB_HISTORY tables are shown on the
following pages.
There have been instances in the company of people who have held the same position more
than once during their tenure with the company. For example, consider the employee Taylor,
who joined the company on 24-MAR-2006. Taylor held the job title SA_REP for the period 24-
MAR-06 to 31-DEC-06 and the job title SA_MAN for the period 01-JAN-07 to 31-DEC-07. Taylor
moved back into the job title of SA_REP, which is his current job title.

DESCRIBE employees
SELECT employee_id, last_name, job_id, hire_date, department_id
FROM employees;

DESCRIBE job_history
SELECT * FROM job_history;
 UNION Operator

A B

The UNION operator returns rows from both queries after eliminating
duplications.

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

The UNION operator returns all rows that are selected by either query. Use the UNION
operator to return all rows from multiple tables and eliminate any duplicate rows.
Guidelines
• The number of columns being selected must be the same.
• The data types of the columns being selected must be in the same data type group
(such as numeric or character).
• The names of the columns need not be identical.
• UNION operates over all of the columns being selected.
• NULL values are not ignored during duplicate checking.
• By default, the output is sorted in ascending order of the columns of the SELECT clause.
 Using the UNION Operator

Display the current and previous job details of all employees.

Display each employee only once.

SELECT employee_id, job_id

FROM employees
UNION
SELECT employee_id, job_id
FROM job_history;

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

The UNION operator eliminates any duplicate records. If records that occur in both the
EMPLOYEES and the JOB_HISTORY tables are identical, the records are displayed only once.
Observe in the output shown in the slide that the record for the employee with the
EMPLOYEE_ID 200 appears twice because the JOB_ID is different in each row.
Consider the following example:
SELECT employee_id, job_id, department_id
FROM employees
UNION
SELECT employee_id, job_id, department_id
FROM job_history;

…
 UNION ALL Operator

A B

The UNION ALL operator returns rows from both queries, including all
duplications.

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

Use the UNION ALL operator to return all rows from multiple queries.
Guidelines
The guidelines for UNION and UNION ALL are the same, with the following two exceptions
that pertain to UNION ALL: Unlike UNION, duplicate rows are not eliminated and the output is
not sorted by default.
 Using the UNION ALL Operator

Display the current and previous departments of all employees.

SELECT employee_id, job_id, department_id
FROM employees
UNION ALL
SELECT employee_id, job_id, department_id
FROM job_history
ORDER BY employee_id;

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

In the example, 30 rows are selected. The combination of the two tables totals to 30 rows.
The UNION ALL operator does not eliminate duplicate rows. UNION returns all distinct rows
selected by either query. UNION ALL returns all rows selected by either query, including all
duplicates. Consider the query in the slide, now written with the UNION clause:
SELECT employee_id, job_id,department_id
FROM employees
UNION
SELECT employee_id, job_id,department_id
FROM job_history
ORDER BY employee_id;
The preceding query returns 29 rows. This is because it eliminates the following row (because
it is a duplicate):
 INTERSECT Operator

A B

The INTERSECT operator returns rows that are common to both queries.

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

Use the INTERSECT operator to return all rows that are common to multiple queries.
Guidelines
• The number of columns and the data types of the columns being selected by the
SELECT statements in the queries must be identical in all the SELECT statements used
in the query. The names of the columns, however, need not be identical.
• Reversing the order of the intersected tables does not alter the result.
• INTERSECT does not ignore NULL values.
 Using the INTERSECT Operator

Display the employee IDs and job IDs of those employees who
currently have a job title that is the same as their previous one
(that is, they changed jobs but have now gone back to doing
the same job they did previously).

SELECT employee_id, job_id

FROM employees
INTERSECT
SELECT employee_id, job_id
FROM job_history;

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

In the example in this slide, the query returns only those records that have the same values in
the selected columns in both tables.
What will be the results if you add the DEPARTMENT_ID column to the SELECT statement
from the EMPLOYEES table and add the DEPARTMENT_ID column to the SELECT statement
from the JOB_HISTORY table, and run this query? The results may be different because of
the introduction of another column whose values may or may not be duplicates.
Example
SELECT employee_id, job_id, department_id
FROM employees
INTERSECT
SELECT employee_id, job_id, department_id
FROM job_history;

Employee 200 is no longer part of the results because the EMPLOYEES.DEPARTMENT_ID

value is different from the JOB_HISTORY.DEPARTMENT_ID value.
 MINUS Operator

A B

The MINUS operator returns all the distinct rows selected by the first query, but
not present in the second query result set.

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

Use the MINUS operator to return all distinct rows selected by the first query, but not present
in the second query result set (the first SELECT statement MINUS the second SELECT
statement).
Note: The number of columns must be the same and the data types of the columns being
selected by the SELECT statements in the queries must belong to the same data type group in
all the SELECT statements used in the query. The names of the columns, however, need not
be identical.
 Using the MINUS Operator

Display the employee IDs of those employees who have not

changed their jobs even once.

SELECT employee_id
FROM employees
MINUS
SELECT employee_id
FROM job_history;

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

In the example in the slide, the employee IDs in the JOB_HISTORY table are subtracted from
those in the EMPLOYEES table. The results set displays the employees remaining after the
subtraction; they are represented by rows that exist in the EMPLOYEES table, but do not exist
in the JOB_HISTORY table. These are the records of the employees who have not changed
their jobs even once.
 Matching the SELECT Statements

• Using the UNION operator, display the location ID,

department name, and the state where it is located.
• You must match the data type (using the TO_CHAR
function or any other conversion functions) when columns
do not exist in one or the other table.

SELECT location_id, department_name "Department",

TO_CHAR(NULL) "Warehouse location"
FROM departments
UNION
SELECT location_id, TO_CHAR(NULL) "Department",
state_province
FROM locations;

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

Because the expressions in the SELECT lists of the queries must match in number, you can
use the dummy columns and the data type conversion functions to comply with this rule. In
the slide, the name, Warehouse location, is given as the dummy column heading. The
TO_CHAR function is used in the first query to match the VARCHAR2 data type of the
state_province column that is retrieved by the second query. Similarly, the TO_CHAR
function in the second query is used to match the VARCHAR2 data type of the
department_name column that is retrieved by the first query.
The output of the query is shown:

…
 Matching the SELECT Statement: Example

Using the UNION operator, display the employee ID, job ID, and
salary of all employees.

SELECT employee_id, job_id,salary

FROM employees
UNION
SELECT employee_id, job_id,0
FROM job_history;

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

The EMPLOYEES and JOB_HISTORY tables have several columns in common (for example,
EMPLOYEE_ID, JOB_ID, and DEPARTMENT_ID). But what if you want the query to display the
employee ID, job ID, and salary using the UNION operator, knowing that the salary exists
only in the EMPLOYEES table?
The code example in the slide matches the EMPLOYEE_ID and JOB_ID columns in the
EMPLOYEES and JOB_HISTORY tables. A literal value of 0 is added to the JOB_HISTORY
SELECT statement to match the numeric SALARY column in the EMPLOYEES SELECT
statement.
In the results shown in the slide, each row in the output that corresponds to a record from the
JOB_HISTORY table contains a 0 in the SALARY column.
 Using the ORDER BY Clause in Set Operations

• The ORDER BY clause can appear only once at the end of

the compound query.
• Component queries cannot have individual ORDER BY
clauses.
• The ORDER BY clause recognizes only the columns of the
first SELECT query.
• By default, the first column of the first SELECT query is
used to sort the output in an ascending order.

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

The ORDER BY clause can be used only once in a compound query. If used, the ORDER BY
clause must be placed at the end of the query. The ORDER BY clause accepts the column
name or an alias. By default, the output is sorted in ascending order in the first column of the
first SELECT query.
Note: The ORDER BY clause does not recognize the column names of the second SELECT
query. To avoid confusion over column names, it is a common practice to ORDER BY column
positions.
For example, in the following statement, the output will be shown in ascending order of
job_id.
SELECT employee_id, job_id,salary
FROM employees
UNION
SELECT employee_id, job_id,0
FROM job_history
ORDER BY 2;
If you omit ORDER BY, by default, the output will be sorted in ascending order of
employee_id. You cannot use the columns from the second query to sort the output.
 Quiz

Identify the two set operator guidelines.

a. The expressions in the SELECT lists must match in
number.
b. Parentheses may not be used to alter the sequence of
execution.
c. The data type of each column in the second query must
match the data type of its corresponding column in the first
query.
d. The ORDER BY clause can be used only once in a
compound query, unless a UNION ALL operator is used.

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

 Summary

In this lesson, you should have learned how to use:

• UNION to return all distinct rows
• UNION ALL to return all rows, including duplicates
• INTERSECT to return all rows that are shared by both
queries
• MINUS to return all distinct rows that are selected by the
first query, but not by the second
• ORDER BY only at the very end of the statement

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

• The UNION operator returns all the distinct rows selected by each query in the
compound query. Use the UNION operator to return all rows from multiple tables and
eliminate any duplicate rows.
• Use the UNION ALL operator to return all rows from multiple queries. Unlike the case
with the UNION operator, duplicate rows are not eliminated and the output is not sorted
by default.
• Use the INTERSECT operator to return all rows that are common to multiple queries.
• Use the MINUS operator to return rows returned by the first query that are not present in
the second query.
• Remember to use the ORDER BY clause only at the very end of the compound
statement.
• Make sure that the corresponding expressions in the SELECT lists match in number and
data type.
 Practice 9: Overview

In this practice, you create reports by using:

• The UNION operator
• The INTERSECT operator
• The MINUS operator

Copyright © 2013, Oracle and/or its affiliates. All rights reserved.

In this practice, you write queries using the set operators.