|Oracle® Database Concepts
11g Release 2 (11.2)
Part Number E16508-04
This chapter explains how integrity constraints enforce the business rules associated with a database and prevent the entry of invalid information into tables.
This chapter contains the following sections:
See Also:"Overview of Tables"
Business rules specify conditions and relationships that must always be true or must always be false. For example, each company defines its own policies about salaries, employee numbers, inventory tracking, and so on. It is important that data maintain data integrity, which is adherence to these rules, as determined by the database administrator or application developer.
When designing a database application, developers have various options for guaranteeing the integrity of data stored in the database. These options include:
Enforcing business rules with triggered stored database procedures, as described in "Overview of Triggers"
Using stored procedures to completely control access to data, as described in "Introduction to Server-Side Programming"
Enforcing business rules in the code of a database application
Using Oracle Database integrity constraints, which are rules defined at the column or object level that restrict values in the database
This chapter explains the basic concepts of integrity constraints.
An integrity constraint is a schema object that is created and dropped using SQL. To enforce data integrity, use integrity constraints unless it is not possible. Advantages of integrity constraints over alternatives for enforcing data integrity include:
Because you define integrity constraints using SQL statements, no additional programming is required when you define or alter a table. The SQL statements are easy to write and eliminate programming errors.
Integrity constraints are defined for tables and are stored in the data dictionary (see "Overview of the Data Dictionary"). Thus, data entered by all applications must adhere to the same integrity constraints. If the rules change at the table level, then applications need not change. Also, applications can use metadata in the data dictionary to immediately inform users of violations, even before the database checks the SQL statement.
Flexibility when loading data
You can disable integrity constraints temporarily to avoid performance overhead when loading large amounts of data. When the data load is complete, you can re-enable the integrity constraints.
Oracle Database enables you to apply constraints both at the table and column level. A constraint specified as part of the definition of a column or attribute is called an inline specification. A constraint specified as part of the table definition is called an out-of-line specification.
The term key is used in the definitions of several types of integrity constraints. A key is the column or set of columns included in the definition of certain types of integrity constraints. Keys describe the relationships between the tables and columns of a relational database. Individual values in a key are called key values.
Table 5-1 describes the types of constraints. Each can be specified either inline or out-of-line, except for
NOT NULL, which must be inline.
Table 5-1 Types of Constraints
|Constraint Type||Description||See Also|
Allows or disallows inserts or updates of rows containing a null in a specified column.
Dictates types of data manipulation allowed on values in a
Oracle Database Object-Relational Developer's Guide to learn about
NOT NULL constraints are intended for columns that must not lack values. For example, the
hr.employees table requires a value in the
last_name column. An attempt to insert an employee row without a last name generates an error:
SQL> INSERT INTO hr.employees (employee_id, last_name) values (999, 'Smith'); . . . ERROR at line 1: ORA-01400: cannot insert NULL into ("HR"."EMPLOYEES"."LAST_NAME")
You can only add a column with a
NOT NULL constraint if the table does not contain any rows or if you specify a default value.
Oracle Database 2 Day Developer's Guide for examples of adding
NOT NULL constraints to a table
Oracle Database SQL Language Reference for restrictions on using
NOT NULL constraints
Oracle Database Advanced Application Developer's Guide to learn when to use the
NOT NULL constraint
A unique key constraint requires that every value in a column or set of columns be unique. No rows of a table may have duplicate values in a column (the unique key) or set of columns (the composite unique key) with a unique key constraint.
Note:The term key refers only to the columns defined in the integrity constraint. Because the database enforces a unique constraint by implicitly creating or reusing an index on the key columns, the term unique key is sometimes incorrectly used as a synonym for unique key constraint or unique index.
Unique key constraints are appropriate for any column where duplicate values are not allowed. Unique constraints differ from primary key constraints, whose purpose is to identify each table row uniquely, and typically contain values that have no significance other than being unique. Examples of unique keys include:
A customer phone number, where the primary key is the customer number
A department name, where the primary key is the department number
As shown in Example 2-1, a unique key constraint exists on the
hr.employees table. The relevant part of the statement is as follows:
CREATE TABLE employees ( ... , email VARCHAR2(25) CONSTRAINT emp_email_nn NOT NULL ... , CONSTRAINT emp_email_uk UNIQUE (email) ... );
emp_email_uk constraint ensures that no two employees have the same email address, as shown in Example 5-1.
Example 5-1 Unique Constraint
SQL> SELECT employee_id, last_name, email FROM employees WHERE email = 'PFAY'; EMPLOYEE_ID LAST_NAME EMAIL ----------- ------------------------- ------------------------- 202 Fay PFAY SQL> INSERT INTO employees (employee_id, last_name, email, hire_date, job_id) 1 VALUES (999,'Fay','PFAY',SYSDATE,'ST_CLERK'); . . . ERROR at line 1: ORA-00001: unique constraint (HR.EMP_EMAIL_UK) violated
NOT NULL constraint is also defined, a null always satisfies a unique key constraint. Thus, columns with both unique key constraints and
NOT NULL constraints are typical. This combination forces the user to enter values in the unique key and eliminates the possibility that new row data conflicts with existing row data.
Note:Because of the search mechanism for unique key constraints on multiple columns, you cannot have identical values in the non-null columns of a partially null composite unique key constraint.
In a primary key constraint, the values in the group of one or more columns subject to the constraint uniquely identify the row. Each table can have one primary key, which in effect names the row and ensures that no duplicate rows exist.
A primary key can be natural or a surrogate. A natural key is a meaningful identifier made of existing attributes in a table. For example, a natural key could be a postal code in a lookup table. In contrast, a surrogate key is a system-generated incrementing identifier that ensures uniqueness within a table. Typically, surrogate keys are generated by a sequence.
The Oracle Database implementation of the primary key constraint guarantees that the following statements are true:
No two rows have duplicate values in the specified column or set of columns.
The primary key columns do not allow nulls.
A typical situation calling for a primary key is the numeric identifier for an employee. Each employee must have a unique ID. A employee must be described by one and only one row in the
Example 5-1 indicates that an existing employee has the employee ID of 202, where the employee ID is the primary key. The following example shows an attempt to add an employee with the same employee ID and an employee with no ID:
SQL> INSERT INTO employees (employee_id, last_name, email, hire_date, job_id) 1 VALUES (202,'Chan','ICHAN',SYSDATE,'ST_CLERK'); . . . ERROR at line 1: ORA-00001: unique constraint (HR.EMP_EMP_ID_PK) violated SQL> INSERT INTO employees (last_name) VALUES ('Chan'); . . . ERROR at line 1: ORA-01400: cannot insert NULL into ("HR"."EMPLOYEES"."EMPLOYEE_ID")
The database enforces primary key constraints with an index. Usually, a primary key constraint created for a column implicitly creates a unique index and a
NOT NULL constraint. Note the following exceptions to this rule:
In some cases, as when you create a primary key with a deferrable constraint, the generated index is not unique.
If a usable index exists when a primary key constraint is created, then the constraint reuses this index and does not implicitly create a new one.
By default the name of the implicitly created index is the name of the primary key constraint. You can also specify a user-defined name for an index. You can specify storage options for the index by including the
ENABLE clause in the
CREATE TABLE or
ALTER TABLE statement used to create the constraint.
See Also:Oracle Database 2 Day Developer's Guide and Oracle Database Advanced Application Developer's Guide to learn how to add primary key constraints to a table
Whenever two tables contain one or more common columns, Oracle Database can enforce the relationship between the two tables through a foreign key constraint, also called a referential integrity constraint. The constraint requires that for each value in the column on which the constraint is defined, the value in the other specified other table and column must match. An example of a referential integrity rule is an employee can work for only an existing department.
Table 5-2 lists terms associated with referential integrity constraints.
Table 5-2 Referential Integrity Constraint Terms
The column or set of columns included in the definition of the constraint that reference a referenced key. For example, the
The value of foreign keys can match either the referenced primary or unique key value, or be null. If any column of a composite foreign key is null, then the non-null portions of the key do not have to match any corresponding portion of a parent key.
Dependent or child table
The table that includes the foreign key. This table is dependent on the values present in the referenced unique or primary key. For example, the
Referenced or parent table
The table that is referenced by the foreign key of the child table. It is this table's referenced key that determines whether specific inserts or updates are allowed in the child table. For example, the
Figure 5-1 shows a foreign key on the
employees.department_id column. It guarantees that every value in this column must match a value in the
departments.department_id column. Thus, no erroneous department numbers can exist in the
Figure 5-1 Referential Integrity Constraints
See Also:Oracle Database 2 Day Developer's Guide and Oracle Database Advanced Application Developer's Guide to learn how to add foreign key constraints to a table
Figure 5-2 shows a self-referential integrity constraint. In this case, a foreign key references a parent key in the same table.
In Figure 5-2, the referential integrity constraint ensures that every value in the
employees.manager_id column corresponds to an existing value in the
employees.employee_id column. For example, the manager for employee 102 must exist in the
employees table. This constraint eliminates the possibility of erroneous employee numbers in the
Figure 5-2 Single Table Referential Constraints
The relational model permits the value of foreign keys to match either the referenced primary or unique key value, or be null. For example, a user could insert a row into
hr.employees without specifying a department ID.
If any column of a composite foreign key is null, then the non-null portions of the key do not have to match any corresponding portion of a parent key.
The relationship between foreign key and parent key has implications for deletion of parent keys. For example, if a user attempts to delete the record for this department, then what happens to the records for employees in this department?
No action on deletion or update
In the normal case, users cannot modify referenced key values if the results would violate referential integrity. For example, if
employees.department_id is a foreign key to
departments, and if employees belong to a particular department, then an attempt to delete the row for this department violates the constraint.
A deletion cascades (
DELETE CASCADE) when rows containing referenced key values are deleted, causing all rows in child tables with dependent foreign key values to also be deleted. For example, the deletion of a row in
departments causes rows for all employees in this department to be deleted.
Deletions that set null
A deletion sets null (
DELETE SET NULL) when rows containing referenced key values are deleted, causing all rows in child tables with dependent foreign key values to set those values to null. For example, the deletion of a department row sets the
department_id column value to null for employees in this department.
Table 5-3 outlines the DML statements allowed by the different referential actions on the key values in the parent table, and the foreign key values in the child table.
Table 5-3 DML Statements Allowed by Update and Delete No Action
|DML Statement||Issued Against Parent Table||Issued Against Child Table|
Always OK if the parent key value is unique
OK only if the foreign key value exists in the parent key or is partially or all null
Allowed if the statement does not leave any rows in the child table without a referenced parent key value
Allowed if the new foreign key value still references a referenced key value
Allowed if no rows in the child table reference the parent key value
Note:Other referential actions not supported by
FOREIGN KEYintegrity constraints of Oracle Database can be enforced using database triggers. See "Overview of Triggers".
See Also:Oracle Database SQL Language Reference to learn about the
As a rule, foreign keys should be indexed. The only exception is when the matching unique or primary key is never updated or deleted. Indexing the foreign keys in child tables provides the following benefits:
Prevents a full table lock on the child table. Instead, the database acquires a row lock on the index.
Removes the need for a full table scan of the child table. As an illustration, assume that a user removes the record for department 10 from the
departments table. If
employees.department_id is not indexed, then the database must scan
employees to see if any employees exist in department 10.
A check constraint on a column or set of columns requires that a specified condition be true or unknown for every row. If DML results in the condition of the constraint evaluating to false, then the SQL statement is rolled back.
The chief benefit of check constraints is the ability to enforce very specific integrity rules. For example, you could use check constraints to enforce the following rules in the
salary column must not have a value greater than 10000.
commission column must have a value that is not greater than the salary.
The following example creates a maximum salary constraint on
employees and demonstrates what happens when a statement attempts to insert a row containing a salary that exceeds the maximum:
SQL> ALTER TABLE employees ADD CONSTRAINT max_emp_sal CHECK (salary < 10001); SQL> INSERT INTO employees (employee_id,last_name,email,hire_date,job_id,salary) 1 VALUES (999,'Green','BGREEN',SYSDATE,'ST_CLERK',20000); . . . ERROR at line 1: ORA-02290: check constraint (HR.MAX_EMP_SAL) violated
A single column can have multiple check constraints that reference the column in its definition. For example, the
salary column could have one constraint that prevents values over 10000 and a separate constraint that prevents values less than 500.
If multiple check constraints exist for a column, then they must be designed so their purposes do not conflict. No order of evaluation of the conditions can be assumed. The database does not verify that check conditions are not mutually exclusive.
See Also:Oracle Database SQL Language Reference to learn about restrictions for check constraints
The database enables you to specify whether a constraint applies to existing data or future data. If a constraint is enabled, then the database checks new data as it is entered or updated. Data that does not conform to the constraint cannot enter the database. For example, enabling a
NOT NULL constraint on
employees.department_id guarantees that every future row has a department ID. If a constraint is disabled, then the table can contain rows that violate the constraint.
You can set constraints to validate (
VALIDATE) or not validate (
NOVALIDATE) existing data. If
VALIDATE is specified, then existing data must conform to the constraint. For example, enabling a
NOT NULL constraint on
employees.department_id and setting it to
VALIDATE checks that every existing row has a department ID. If
NOVALIDATE is specified, then existing data need not conform to the constraint.
The behavior of
NOVALIDATE always depends on whether the constraint is enabled or disabled. Table 5-4 summarizes the relationships.
Table 5-4 Checks on Modified and Existing Data
|Modified Data||Existing Data||Summary|
Existing and future data must obey the constraint. An attempt to apply a new constraint to a populated table results in an error if existing rows violate the constraint.
The database checks the constraint, but it need not be true for all rows. Thus, existing rows can violate the constraint, but new or modified rows must conform to the rules.
The database disables the constraint, drops its index, and prevents modification of the constrained columns.
The constraint is not checked and is not necessarily true.
See Also:Oracle Database SQL Language Reference to learn about constraint states
Every constraint is either in a not deferrable (default) or deferrable state. This state determines when Oracle Database checks the constraint for validity. The following graphic depicts the options for deferrable constraints.
If a constraint is not deferrable, then Oracle Database never defers the validity check of the constraint to the end of the transaction. Instead, the database checks the constraint at the end of each statement. If the constraint is violated, then the statement rolls back.
For example, assume that you create a nondeferrable
NOT NULL constraint for the
employees.last_name column. If a user attempts to insert a row with no last name, then the database immediately rolls back the statement because the
NOT NULL constraint is violated. No row is inserted.
A deferrable constraint permits a transaction to use the
SET CONSTRAINT clause to defer checking of this constraint until a
COMMIT statement is issued. If you make changes to the database that might violate the constraint, then this setting effectively lets you disable the constraint until all the changes are complete.
The database checks the constraint immediately after each statement executes. If the constraint is violated, then the database rolls back the statement.
The database checks the constraint when a
COMMIT is issued. If the constraint is violated, then the database rolls back the transaction.
Assume that a deferrable
NOT NULL constraint on
employees.last_name is set to
INITIALLY DEFERRED. A user creates a transaction with 100
INSERT statements, some of which have null values for
last_name. When the user attempts to commit, the database rolls back all 100 statements. However, if this constraint were set to
INITIALLY IMMEDIATE, then the database would not roll back the transaction.
If a constraint causes an action, then the database considers this action as part of the statement that caused it, whether the constraint is deferred or immediate. For example, deleting a row in
departments causes the deletion of all rows in
employees that reference the deleted department row. In this case, the deletion from
employees is considered part of the
DELETE statement executed against
See Also:Oracle Database SQL Language Reference for information about constraint attributes and their default values
employees table has the structure shown in Figure 5-2.
The self-referential constraint makes entries in the
manager_id column dependent on the values of the
Consider the insertion of the first row into the
employees table. No rows currently exist, so how can a row be entered if the value in the
manager_id column cannot reference any existing value in the
employee_id column? Some possibilities are:
A null can be entered for the
manager_id column of the first row, if the
manager_id column does not have a
NULL constraint defined on it.
Because nulls are allowed in foreign keys, this row is inserted into the table.
The same value can be entered in the
manager_id columns, specifying that the employee is his or her own manager.
This case reveals that Oracle Database performs its constraint checking after the statement has been completely run. To allow a row to be entered with the same values in the parent key and the foreign key, the database must first run the statement (that is, insert the new row) and then determine whether any row in the table has an
employee_id that corresponds to the
manager_id of the new row.
A multiple row
INSERT statement, such as an
INSERT statement with nested
SELECT statement, can insert rows that reference one another.
For example, the first row might have 200 for employee ID and 300 for manager ID, while the second row has 300 for employee ID and 200 for manager. Constraint checking is deferred until the complete execution of the statement. All rows are inserted first, and then all rows are checked for constraint violations.
Consider the same self-referential integrity constraint in a different scenario. The company has been sold. Because of this sale, all employee numbers must be updated to be the current value plus 5000 to coordinate with the employee numbers of the new company. Because manager numbers are really employee numbers (see Figure 5-3), the manager numbers must also increase by 5000.
Figure 5-3 The employees Table Before Updates
You could execute the following SQL statement to update the values:
UPDATE employees SET employee_id = employee_id + 5000, manager_id = manager_id + 5000;
Although a constraint is defined to verify that each
manager_id value matches an
employee_id value, the preceding statement is legal because the database effectively checks constraints after the statement completes. Figure 5-4 shows that the database performs the actions of the entire SQL statement before checking constraints.
Figure 5-4 Constraint Checking
The examples in this section illustrate the constraint checking mechanism during
UPDATE statements, but the database uses the same mechanism for all types of DML statements. The same mechanism is used for all types of constraints, not just self-referential constraints.