Oracle sql joins

Editor's Notes

• #2: Schedule:TimingTopic 55 minutesLecture 55 minutesPractice 110 minutesTotal
• #3: Lesson Aim This lesson covers how to obtain data from more than one table.
• #4: Data from Multiple Tables Sometimes you need to use data from more than one table. In the slide example, the report displays data from two separate tables. Employee IDs exist in the EMPLOYEES table. Department IDs exist in both the EMPLOYEES and DEPARTMENTS tables. Location IDs exist in the DEPARTMENTS table. To produce the report, you need to link the EMPLOYEES and DEPARTMENTS tables and access data from both of them.
• #5: Cartesian Products When a join condition is invalid or omitted completely, the result is a Cartesian product, in which all combinations of rows are displayed. All rows in the first table are joined to all rows in the second table. A Cartesian product tends to generate a large number of rows, and the result is rarely useful. You should always include a valid join condition in a WHERE clause, unless you have a specific need to combine all rows from all tables. Cartesian products are useful for some tests when you need to generate a large number of rows to simulate a reasonable amount of data.
• #6: Cartesian Products (continued) A Cartesian product is generated if a join condition is omitted. The example on the slide displays employee last name and department name from the EMPLOYEES and DEPARTMENTS tables. Because no WHERE clause has been specified, all rows (20 rows) from the EMPLOYEES table are joined with all rows (8 rows) in the DEPARTMENTS table, thereby generating 160 rows in the output. SELECT last_name, department_name dept_name FROM employees, departments; Instructor Note Demo: 4_cart.sql Purpose: To illustrate executing a Cartesian product
• #7: Types of Joins The Oracle9i database offers join syntax that is SQL: 1999 compliant. Prior to the 9i release, the join syntax was different from the ANSI standards. The new SQL: 1999 compliant join syntax does not offer any performance benefits over the Oracle proprietary join syntax that existed in prior releases. Instructor Note Do not get into details of all the types of joins now. Explain each join one by one, as is done in the following slides.
• #8: Defining Joins When data from more than one table in the database is required, a join condition is used. Rows in one table can be joined to rows in another table according to common values existing in corresponding columns, that is, usually primary and foreign key columns. To display data from two or more related tables, write a simple join condition in the WHERE clause. In the syntax: table1.columndenotes the table and column from which data is retrieved table1.column1 =is the condition that joins (or relates) the tables togethertable2.column2 Guidelines When writing a SELECT statement that joins tables, precede the column name with the table name for clarity and to enhance database access. If the same column name appears in more than one table, the column name must be prefixed with the table name. To join n tables together, you need a minimum of n-1 join conditions. For example, to join four tables, a minimum of three joins is required. This rule may not apply if your table has a concatenated primary key, in which case more than one column is required to uniquely identify each row. For more information, see Oracle9i SQL Reference, “SELECT.”
• #9: Equijoins To determine an employee’s department name, you compare the value in the DEPARTMENT_ID column in the EMPLOYEES table with the DEPARTMENT_ID values in the DEPARTMENTS table. The relationship between the EMPLOYEES and DEPARTMENTS tables is an equijoin—that is, values in the DEPARTMENT_ID column on both tables must be equal. Frequently, this type of join involves primary and foreign key complements. Note: Equijoins are also called simple joins or inner joins. Instructor Note Explain the use of a decision matrix for simplifying writing joins. For example, if you want to display the name and department number of all the employees who are in the same department as Goyal, you can start by making the following decision tree: Columns to DisplayOriginating TableCondition last_nameemployeeslast_name='Goyal' department_namedepartmentsemployees.department_id = departments.department_id Now the SQL statement can be easily formulated by looking at the decision matrix. The first column gives the column list in the SELECT statement, the second column gives the tables for the FROM clause, and the third column gives the condition for the WHERE clause.
• #10: Retrieving Records with Equijoins In the slide example: The SELECT clause specifies the column names to retrieve: employee last name, employee number, and department number, which are columns in the EMPLOYEES table department number, department name, and location ID, which are columns in the DEPARTMENTS table The FROM clause specifies the two tables that the database must access: EMPLOYEES table DEPARTMENTS table The WHERE clause specifies how the tables are to be joined: EMPLOYEES.DEPARTMENT_ID = DEPARTMENTS.DEPARTMENT_ID Because the DEPARTMENT_ID column is common to both tables, it must be prefixed by the table name to avoid ambiguity.
• #11: Qualifying Ambiguous Column Names You need to qualify the names of the columns in the WHERE clause with the table name to avoid ambiguity. Without the table prefixes, the DEPARTMENT_ID column could be from either the DEPARTMENTS table or the EMPLOYEES table. It is necessary to add the table prefix to execute your query. If there are no common column names between the two tables, there is no need to qualify the columns. However, using the table prefix improves performance, because you tell the Oracle Server exactly where to find the columns. The requirement to qualify ambiguous column names is also applicable to columns that may be ambiguous in other clauses, such as the SELECT clause or the ORDER BY clause. Instructor Note Demo: 4_loc.sql Purpose: To illustrate a SELECT clause with no aliases.
• #12: Table Aliases Qualifying column names with table names can be very time consuming, particularly if table names are lengthy. You can use table aliases instead of table names. Just as a column alias gives a column another name, a table alias gives a table another name. Table aliases help to keep SQL code smaller, therefore using less memory. Notice how table aliases are identified in the FROM clause in the example. The table name is specified in full, followed by a space and then the table alias. The EMPLOYEES table has been given an alias of e, and the DEPARTMENTS table has an alias of d. Guidelines Table aliases can be up to 30 characters in length, but shorter is better. If a table alias is used for a particular table name in the FROM clause, then that table alias must be substituted for the table name throughout the SELECT statement. Table aliases should be meaningful. The table alias is valid only for the current SELECT statement.
• #13: Additional Search Conditions Sometimes you may need to join more than two tables. For example, to display the last name, the department name, and the city for each employee, you have to join the EMPLOYEES, DEPARTMENTS, and LOCATIONS tables. SELECT e.last_name, d.department_name, l.city FROM employees e, departments d, locations l WHERE e.department_id = d.department_id AND d.location_id = l.location_id;
• #14: Non-Equijoins A non-equijoin is a join condition containing something other than an equality operator. The relationship between the EMPLOYEES table and the JOB_GRADES table has an example of a non-equijoin. A relationship between the two tables is that the SALARY column in the EMPLOYEES table must be between the values in the LOWEST_SALARY and HIGHEST_SALARY columns of the JOB_GRADES table. The relationship is obtained using an operator other than equals (=).
• #15: Non-Equijoins (continued) The slide example creates a non-equijoin to evaluate an employee’s salary grade. The salary must be between any pair of the low and high salary ranges. It is important to note that all employees appear exactly once when this query is executed. No employee is repeated in the list. There are two reasons for this: None of the rows in the job grade table contain grades that overlap. That is, the salary value for an employee can lie only between the low salary and high salary values of one of the rows in the salary grade table. All of the employees’ salaries lie within the limits provided by the job grade table. That is, no employee earns less than the lowest value contained in the LOWEST_SAL column or more than the highest value contained in the HIGHEST_SAL column. Note: Other conditions, such as <= and >= can be used, but BETWEEN is the simplest. Remember to specify the low value first and the high value last when using BETWEEN. Table aliases have been specified in the slide example for performance reasons, not because of possible ambiguity. Instructor Note Explain that BETWEEN … AND … is actually translated by the Oracle server to a pair of AND conditions (a >= lower limit) and (a <= higher limit) and IN ( … ) is translated by the Oracle server to a set of OR conditions (a = value1 OR a = value2 OR a = value3 ). So using BETWEEN … AND … , IN(…) has no performance benefits; the benefit is logical simplicity.
• #16: Returning Records with No Direct Match with Outer Joins If a row does not satisfy a join condition, the row will not appear in the query result. For example, in the equijoin condition of EMPLOYEES and DEPARTMENTS tables, employee Grant does not appear because there is no department ID recorded for her in the EMPLOYEES table. Instead of seeing 20 employees in the result set, you see 19 records. SELECT e.last_name, e.department_id, d.department_name FROM employees e, departments d WHERE e.department_id = d.department_id;
• #17: Using Outer Joins to Return Records with No Direct Match The missing rows can be returned if an outer join operator is used in the join condition. The operator is a plus sign enclosed in parentheses (+), and it is placed on the “side” of the join that is deficient in information. This operator has the effect of creating one or more null rows, to which one or more rows from the nondeficient table can be joined. In the syntax: table1.column =is the condition that joins (or relates) the tables together. table2.column (+)is the outer join symbol, which can be placed on either side of theWHERE clause condition, but not on both sides. (Place the outerjoin symbol following the name of the column in the table without the matching rows.) Instructor Note Demo: 4_ejoin.sql Purpose: To illustrate an equijoin leading to an outer join.
• #18: Using Outer Joins to Return Records with No Direct Match (continued) The slide example displays employee last names, department ID’s and department names. The Contracting department does not have any employees. The empty value is shown in the output shown. Outer Join Restrictions The outer join operator can appear on only one side of the expression—the side that has information missing. It returns those rows from one table that have no direct match in the other table. A condition involving an outer join cannot use the IN operator or be linked to another condition by the OR operator. Instructor Note The UNION operator works around the issue of being able to use an outer join operator on one side of the expression. The ANSI full outer join also allows you to have an outer join on both sides of the expression. It is discussed later in this lesson. Demo: 4_ojoin.sql Purpose: To illustrate an outer join.
• #19: Joining a Table to Itself Sometimes you need to join a table to itself. To find the name of each employee’s manager, you need to join the EMPLOYEES table to itself, or perform a self join. For example, to find the name of Whalen’s manager, you need to: Find Whalen in the EMPLOYEES table by looking at the LAST_NAME column. Find the manager number for Whalen by looking at the MANAGER_ID column. Whalen’s manager number is 101. Find the name of the manager with EMPLOYEE_ID 101 by looking at the LAST_NAME column. Kochhar’s employee number is 101, so Kochhar is Whalen’s manager. In this process, you look in the table twice. The first time you look in the table to find Whalen in the LAST_NAME column and MANAGER_ID value of 101. The second time you look in the EMPLOYEE_ID column to find 101 and the LAST_NAME column to find Kochhar. Instructor Note Show the data from the EMPLOYEES table and point out how each manager is also an employee.
• #20: Joining a Table to Itself (continued) The slide example joins the EMPLOYEES table to itself. To simulate two tables in the FROM clause, there are two aliases, namely w and m, for the same table, EMPLOYEES. In this example, the WHERE clause contains the join that means “where a worker’s manager number matches the employee number for the manager.” Instructor Note Point out the following to the students: The column heading in the result of the query on the slide seems meaningless. A meaningful column alias should have been used instead. There are only 19 rows in the output, but there are 20 rows in the EMPLOYEES table. This occurs because employee King, who is the president, does not have a manager.
• #21: Creating Cross Joins The example on the slide gives the same results as the following: SELECT last_name, department_name FROM employees, departments;
• #22: Joins - Comparing SQL: 1999 to Oracle Syntax
• #23: Example of LEFT OUTER JOIN This query retrieves all rows in the EMPLOYEES table, which is the left table even if there is no match in the DEPARTMENTS table. This query was completed in earlier releases as follows: SELECT e.last_name, e.department_id, d.department_name FROM employees e, departments d WHERE d.department_id (+) = e.department_id;
• #24: Example of RIGHT OUTER JOIN This query retrieves all rows in the DEPARTMENTS table, which is the right table even if there is no match in the EMPLOYEES table. This query was completed in earlier releases as follows: SELECT e.last_name, e.department_id, d.department_name FROM employees e, departments d WHERE d.department_id = e.department_id (+);
• #25: Example of FULL OUTER JOIN This query retrieves all rows in the EMPLOYEES table, even if there is no match in the DEPARTMENTS table. It also retrieves all rows in the DEPARTMENTS table, even if there is no match in the EMPLOYEES table. Instructor Note It was not possible to complete this in earlier releases using outer joins. However, you could accomplish the same results using the UNION operator. SELECT e.last_name, e.department_id, d.department_name FROM employees e, departments d WHERE e.department_id (+) = d.department_id UNION SELECT e.last_name, e.department_id, d.department_name FROM employees e, departments d WHERE e.department_id = d.department_id (+);
• #26: Applying Additional Conditions You can apply additional conditions in the WHERE clause. The example shown performs a join on the EMPLOYEES and DEPARTMENTS tables, and, in addition, displays only employees with a manager ID equal to 149.