Execute the following statement to return nine rows describing the managers of each employee in these departments: select e.last_name employee, e.employee_id, e.manager_id, m.last_name m
Trang 1where d.department_name like 'P%';
Query 2: select d.department_name from departments d
join locations l on
(l.LOCATION_ID=d.LOCATION_ID and d.department_name like 'P%');
Query 1 uses a WHERE clause to restrict the 27 rows created by equijoining the DEPARTMENTS and LOCATIONS tables based on their LOCATION_ID values to the three that contain DEPARTMENT_ID values beginning with the letter “P.” Query 2 implements the condition within the brackets of the ON subclause and returns the same three rows
Five tables are joined in Figure 12-7, resulting in a list describing the top earning employees and geographical information about their departments
EXAM TIP There are three natural join formats The pure natural join uses
the NATURAL JOIN clause and joins two tables based on all columns with shared names The other two formats use the JOIN USING and JOIN
ON clauses and are also referred to as natural joins They do not use the NATURAL keyword
Nonequijoins
Nonequijoins match column values from different tables based on an inequality
expression The value of the join column in each row in the source table is compared
to the corresponding values in the target table A match is found if the expression used in the join, based on an inequality operator, evaluates to true When such a join
is constructed, a nonequijoin is performed.
A nonequijoin is specified using the JOIN ON syntax, but the join condition
contains an inequality operator instead of an equal sign
The format of the syntax for a nonequijoin clause is as follows:
SELECT table1.column, table2.column
FROM table1
[JOIN table2 ON (table1.expr1< table2.expr2)]|
[JOIN table2 ON (table1.expr1 > table2.expr2)]|
[JOIN table2 ON (table1.expr1 <= table2.expr2)]|
[JOIN table2 ON (table1.expr1 >= table2.expr2)]|
[JOIN table2 ON (table1.expr1 BETWEEN table2.expr2 AND table2.expr3)]| [JOIN table2 ON (table1.expr1 LIKE table2 expr2)]
Trang 2Chapter 12: SQL Joins
497
Consider the first 15 rows returned by the query in Figure 12-8 The EMPLOYEES
table is nonequijoined to the JOBS table based on the inequality join condition (2*E
.SALARY < J.MAX_SALARY) The JOBS table stores the salary range for different jobs
in the organization The SALARY value for each employee record is doubled and
compared with all MAX_SALARY values in the JOBS table If the join condition
evaluates to true, the row is returned
The first two rows display the employee with a LAST_NAME of Abel who currently
has a JOB_ID value of SA_REP and earns a SALARY of 11000 These are the only two
rows in the JOBS table that satisfy the inequality join condition (2*E.SALARY < J.MAX_
SALARY) for this employee record
TIP Nonequijoins are not commonly used The BETWEEN range operator
often appears with nonequijoin conditions, since it is simpler to use one
BETWEEN operator in a condition than two nonequijoin conditions based
on (<=) and (>=) operators
Figure 12-7 N-way joins and additional join conditions
Trang 3Join a Table to Itself Using a Self-Join
Storing hierarchical data in a single relational table is accomplished by allocating at least two columns per row One column stores an identifier of the row’s parent record and the second stores the row’s identifier Associating rows with each other based on a
hierarchical relationship requires Oracle to self-join a table to itself.
Joining a Table to Itself Using the JOIN ON Clause
Suppose there is a need to store a family tree in a relational table There are several approaches one could take One option is to use a table called FAMILY with columns named ID, NAME, MOTHER_ID, and FATHER_ID, where each row stores a person’s name, unique ID number, and the ID values for their parents
When two tables are joined, each row from the source table is subjected to the join condition with rows from the target table If the condition evaluates to true, then the joined row, consisting of columns from both tables, is returned
When the join columns originate from the same table, a self-join is required Conceptually, the source table is duplicated to create the target table The self-join works like a regular join between these tables Note that, internally, Oracle does not
Figure 12-8 Nonequijoins
Trang 4Chapter 12: SQL Joins
499
duplicate the table and this description is merely provided to explain the concept of
self-joining Consider the following three queries:
Query 1: select id, name, father_id from family;
Query 2: select name from family where id=&father_id;
Query 3: select f1.name Dad, f2.name Child
from family f1 join family f2 on (f1.id=f2.father_id)
To identify someone’s father in the FAMILY table, you could use query 1 to get
their ID, NAME, and FATHER_ID In query 2, the FATHER_ID value obtained from
the first query can be substituted to obtain the father’s NAME value Notice that both
queries 1 and 2 source information from the FAMILY table
Query 3 performs a self-join with the JOIN ON clause by aliasing the FAMILY
table as f1 and f2 Oracle treats these as different tables even though they point to the
same physical table The first occurrence of the FAMILY table, aliased as f1, is designated
as the source table, while the second, aliased as f2, is assigned as the target table The join
condition in the ON clause is of the format source.child_id=target.parent_id Figure 12-9
shows a sample of FAMILY data and demonstrates a three-way self-join to the same table
Figure 12-9 Self-join
Trang 5table as E and the second instance of the EMPLOYEES table as M Sort the results based on the DEPARTMENT_ID column
1 Start SQL Developer or SQL*Plus and connect to the HR schema
2 Execute the following statement to return nine rows describing the managers
of each employee in these departments:
select e.last_name employee, e.employee_id, e.manager_id, m.last_name manager, e.department_id
from employees e join employees m on (e.manager_id=m.employee_id) where e.department_id in (10,20,30)
order by e.department_id;
View Data That Does Not Meet a Join Condition
by Using Outer Joins
Equijoins match rows between two tables based on the equality of the terms involved
in the join condition Nonequijoins rely on matching rows between tables based on a join condition containing an inequality operator Target table rows with no matching join column in the source table are usually not required When they are required,
however, an outer join is used to fetch them Several variations of outer joins may be
used, depending on whether join column data is missing from the source or target tables or both These outer join techniques are described in the following topics:
• Inner versus outer joins
• Left outer joins
• Right outer joins
• Full outer joins
Inner Versus Outer Joins
When equijoins and nonequijoins are performed, rows from the source and target tables are matched using a join condition formulated with equality and inequality
operators, respectively These are referred to as inner joins An outer join is performed
when rows, that are not retrieved by an inner join, are returned
Two tables sometimes share a master-detail or parent-child relationship In the HR schema the DEPARTMENTS table stores a master list of DEPARTMENT_NAME and
DEPARTMENT_ID values Each EMPLOYEES record has a DEPARTMENT_ID column constrained to be either a value that exists in the DEPARTMENTS table or null This leads to one of the following three scenarios The fourth scenario could occur if the constraint between the tables was removed
Trang 6Chapter 12: SQL Joins
501
1 An employee row has a DEPARTMENT_ID value that matches a row in the
DEPARTMENTS table
2 An employee row has a null value in its DEPARTMENT_ID column
3 There are rows in the DEPARTMENTS table with DEPARTMENT_ID values
that are not stored in any employee records
4 An employee row has a DEPARTMENT_ID value that is not featured in the
DEPARTMENTS table
Rows matching the first scenario are retrieved using a natural inner join between
the two tables The second and third scenarios cause many problems, as these rows
are excluded by inner joins An outer join can be used to include these orphaned rows
in the results set The fourth scenario should rarely occur in a well-designed database,
because foreign key constraints would prevent the insertion of child records with no
parent values Since this row will be excluded by an inner join, it may be retrieved
using an outer join
A left outer join between the source and target tables returns the results of an inner
join as well as rows from the source table excluded by that inner join A right outer
join between the source and target tables returns the results of an inner join as well as
rows from the target table excluded by that inner join If a join returns the results of
an inner join as well as rows from both the source and target tables excluded by that
inner join, then a full outer join has been performed.
Left Outer Joins
The format of the syntax for the LEFT OUTER JOIN clause is as follows:
SELECT table1.column, table2.column
FROM table1
LEFT OUTER JOIN table2
ON (table1.column = table2.column);
A left outer join performs an inner join of table1 and table2 based on the condition
specified after the ON keyword Any rows from the table on the left of the JOIN
keyword excluded for not fulfilling the join condition are also returned Consider
the following two queries:
Query 1: select e.employee_id, e.department_id EMP_DEPT_ID,
d.department_id DEPT_DEPT_ID, d.department_name
from departments d left outer join employees e
on (d.DEPARTMENT_ID=e.DEPARTMENT_ID)
where d.department_name like 'P%';
Query 2: select e.employee_id, e.department_id EMP_DEPT_ID,
d.department_id DEPT_DEPT_ID, d.department_name
from departments d join employees e
on (d.DEPARTMENT_ID=e.DEPARTMENT_ID)
where d.department_name like 'P%';
Trang 7not include this row, since no employees are currently assigned to the department.
A left outer join is shown in Figure 12-10 The inner join produces 27 rows with matching LOCATION_ID values in both tables There are 43 rows in total, which
implies that 16 rows were retrieved from the LOCATIONS table, which is on the left
of the JOIN keyword None of the rows from the DEPARTMENTS table contain any of these 16 LOCATION_ID values
Figure 12-10 Left outer join
Trang 8Chapter 12: SQL Joins
503
Right Outer Joins
The format of the syntax for the RIGHT OUTER JOIN clause is as follows:
SELECT table1.column, table2.column
FROM table1
RIGHT OUTER JOIN table2
ON (table1.column = table2.column);
A right outer join performs an inner join of table1 and table2 based on the join
condition specified after the ON keyword Rows from the table to the right of the JOIN
keyword, excluded by the join condition, are also returned Consider the following
query:
select e.last_name, d.department_name from departments d
right outer join employees e
on (e.department_id=d.department_id)
where e.last_name like 'G%';
The inner join produces seven rows containing details for the employees with
LAST_NAME values that begin with “G.” The EMPLOYEES table is to the right of the
JOIN keyword Any employee records that do not conform to the join condition are
included, provided they conform to the WHERE clause condition In addition, the
right outer join fetches one EMPLOYEE record with a LAST_NAME of Grant This
record currently has a null DEPARTMENT_ID value The inner join excludes the
record, since no DEPARTMENT_ID is assigned to this employee
A right outer join between the JOB_HISTORY and EMPLOYEES tables is shown in
Figure 12-11 The EMPLOYEES table is on the right of the JOIN keyword The DISTINCT
keyword eliminates duplicate combinations of JOB_ID values from the tables The
results show the jobs that employees have historically left The jobs that no employees
have left are also returned
EXAM TIP There are three types of outer join formats Each of them
performs an inner join before including rows the join condition excluded
If a left outer join is performed, then rows excluded by the inner join, to the
left of the JOIN keyword, are also returned If a right outer join is performed,
then rows excluded by the inner join, to the right of the JOIN keyword, are
returned as well
Full Outer Joins
The format of the syntax for the FULL OUTER JOIN clause is as follows:
SELECT table1.column, table2.column
FROM table1
FULL OUTER JOIN table2
ON (table1.column = table2.column);
A full outer join returns the combined results of a left and right outer join An inner
join of table1 and table2 is performed before rows excluded by the join condition from
both tables are merged into the results set
Trang 9The traditional Oracle join syntax does not support a full outer join, which is typically performed by combining the results from left and right outer joins using
the UNION set operator described in Chapter 13 Consider the full outer join shown
in Figure 12-12 The WHERE clause restricting the results to rows with NULL
DEPARTMENT_ ID values shows the orphan rows in both tables There is one record
in the EMPLOYEES table that has no DEPARTMENT_ID values, and there are 16 departments to which no employees belong
Figure 12-11 Right outer join
Trang 10Chapter 12: SQL Joins
505
Generate a Cartesian Product
of Two or More Tables
A Cartesian product of two tables is created by joining each row of the source table with
every row in the target table The number of rows in the result set created by a Cartesian
product is equal to the number of rows in the source table multiplied by the number
of rows in the target table Cartesian products may be formed intentionally using the
ANSI SQL:1999 cross join syntax This technique is described in the next section.
Figure 12-12 Full outer join