Using the set operators

Editor's Notes

• #3 Set Operators The set operators combine the results of two or more component queries into one result. Queries containing set operators are called compound queries. 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.
• #4 Tables Used in This Lesson Two tables are used in this lesson. They are the EMPLOYEES table and the JOB_HISTORY table. The EMPLOYEES table stores the employee details. For the human resource records, this table stores a unique identification number and e-mail address for each employee. The details of the employee’s job identification number, salary, and manager are also stored. Some of the employees earn a commission in addition to their salary; this information is tracked, too. The company organizes the roles of employees into jobs. 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 identification number, 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.
• #5 Tables Used in This Lesson (continued) 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-1998. Taylor held the job title SA_REP for the period 24-MAR-98 to 31-DEC-98 and the job title SA_MAN for the period 01-JAN-99 to 31-DEC-99. Taylor moved back into the job title of SA_REP, which is his current job title. Similarly, consider the employee Whalen, who joined the company on 17-SEP-1987. Whalen held the job title AD_ASST for the period 17-SEP-87 to 17-JUN-93 and the job title AC_ACCOUNT for the period 01-JUL-94 to 31-DEC-98. Whalen moved back into the job title of AD_ASST, which is his current job title. DESCRIBE employees
• #6 Tables Used in This Lesson (continued) SELECT employee_id, last_name, job_id, hire_date, department_id FROM employees; DESCRIBE job_history
• #7 Tables Used in This Lesson (continued) SELECT * FROM job_history;
• #8 UNION Operator 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 and the data types of the columns being selected must be identical in all the SELECT statements used in the query. 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. The IN operator has a higher precedence than the UNION operator. By default, the output is sorted in ascending order of the first column of the SELECT clause.
• #9 Using the UNION Operator 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 on 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;
• #10 Using the UNION Operator (continued) In the preceding output, employee 200 appears three times. Why? Notice the DEPARTMENT_ID values for employee 200. One row has a DEPARTMENT_ID of 90, another 10, and the third 90. Because of these unique combinations of job IDs and department IDs, each row for employee 200 is unique and therefore not considered to be a duplicate. Observe that the output is sorted in ascending order of the first column of the SELECT clause (in this case, EMPLOYEE_ID).
• #11 UNION ALL Operator 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. The DISTINCT keyword cannot be used.
• #12 UNION ALL Operator (continued) 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 on 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 (as it is a duplicate):
• #13 INTERSECT Operator 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 need not be identical. Reversing the order of the intersected tables does not alter the result. INTERSECT does not ignore NULL values.
• #14 INTERSECT Operator (continued) In the example in this slide, the query returns only the 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.
• #15 MINUS Operator Use the MINUS operator to return rows returned by the first query that are not present in the second query (the first SELECT statement MINUS the second SELECT statement). 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 need not be identical. All of the columns in the WHERE clause must be in the SELECT clause for the MINUS operator to work.
• #16 MINUS Operator (continued) In the example in the slide, the employee IDs and job 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.
• #17 Set Operator Guidelines 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 in their WHERE clause must have the same number and type of columns in their SELECT list. For example: SELECT employee_id, department_id FROM employees WHERE (employee_id, department_id) IN (SELECT employee_id, department_id FROM employees UNION SELECT employee_id, department_id FROM job_history); The ORDER BY clause: Can appear only at the very end of the statement Will accept the column name, an alias, or the positional notation The column name or alias, if used in an ORDER BY clause, must be from the first SELECT list. Set operators can be used in subqueries.
• #18 The Oracle Server and Set Operators 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 data type CHAR, the returned values have data type CHAR. If either or both of the queries select values of data type VARCHAR2, the returned values have data type VARCHAR2.
• #19 Matching the SELECT Statements Because the expressions in the select lists of the queries must match in number, you can use dummy columns and the data type conversion functions to comply with this rule. In the slide, the name location is given as the dummy column heading. The TO_NUMBER function is used in the first query to match the NUMBER data type of the LOCATION_ID column retrieved by the second query. Similarly, the TO_DATE function in the second query is used to match the DATE data type of the HIRE_DATE column retrieved by the first query.
• #20 Matching the SELECT Statement: Example 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 preceding results, each row in the output that corresponds to a record from the JOB_HISTORY table contains a 0 in the SALARY column.
• #21 Controlling the Order of Rows By default, the output is sorted in ascending order on the first column. You can use the ORDER BY clause to change this. 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. Without the ORDER BY clause, the code example in the slide produces the following output in the alphabetical order of the first column: Note: Consider a compound query where the UNION set operator is used more than once. In this case, the ORDER BY clause can use only positions rather than explicit expressions. The iSQL*Plus COLUMN Command You can use the iSQL*Plus COLUMN command to customize column headings.