Hibernate.orgCommunity Documentation

Hibernate Validator

JSR 303 Reference Implementation

Reference Guide

4.0.2.GA


Preface
1. Getting started
1.1. Setting up a new Maven project
1.2. Applying constraints
1.3. Validating constraints
1.4. Where to go next?
2. Validation step by step
2.1. Defining constraints
2.1.1. Field-level constraints
2.1.2. Property-level constraints
2.1.3. Class-level constraints
2.1.4. Constraint inheritance
2.1.5. Object graphs
2.2. Validating constraints
2.2.1. Obtaining a Validator instance
2.2.2. Validator methods
2.2.3. ConstraintViolation methods
2.2.4. Message interpolation
2.3. Validating groups
2.3.1. Group sequences
2.3.2. Redefining the default group sequence of a class
2.4. Built-in constraints
3. Creating custom constraints
3.1. Creating a simple constraint
3.1.1. The constraint annotation
3.1.2. The constraint validator
3.1.3. The error message
3.1.4. Using the constraint
3.2. Constraint composition
4. XML configuration
4.1. validation.xml
4.2. Mapping constraints
5. Bootstrapping
5.1. Configuration and ValidatorFactory
5.2. ValidationProviderResolver
5.3. MessageInterpolator
5.4. TraversableResolver
5.5. ConstraintValidatorFactory
6. Integration with other frameworks
6.1. Database schema-level validation
6.2. ORM integration
6.2.1. Hibernate event-based validation
6.2.2. JPA
6.3. Presentation layer validation
7. Further reading

Validating data is a common task that occurs throughout any application, from the presentation layer to the persistence layer. Often the same validation logic is implemented in each layer, proving time consuming and error-prone. To avoid duplication of these validations in each layer, developers often bundle validation logic directly into the domain model, cluttering domain classes with validation code which is really metadata about the class itself.

JSR 303 - Bean Validation - defines a metadata model and API for entity validation. The default metadata source is annotations, with the ability to override and extend the meta-data through the use of XML. The API is not tied to a specific application tier or programming model. It is specifically not tied to either the web tier or the persistence tier, and is available for both server-side application programming, as well as rich client Swing application developers.

Hibernate Validator is the reference implementation of this JSR.

This chapter will show you how to get started with Hibernate Validator, the reference implementation (RI) of Bean Validation. For the following quickstart you need:

To perform a validation of these constraints, we use a Validator instance. Let's have a look at the CarTest class:

Example 1.4. Class CarTest showing validation examples

package com.mycompany;

import static org.junit.Assert.*;

import java.util.Set;

import javax.validation.ConstraintViolation;
import javax.validation.Validation;
import javax.validation.Validator;
import javax.validation.ValidatorFactory;

import org.junit.BeforeClass;
import org.junit.Test;

public class CarTest {

    private static Validator validator;

    @BeforeClass
    public static void setUp() {
        ValidatorFactory factory = Validation.buildDefaultValidatorFactory();
        validator = factory.getValidator();
    }

    @Test
    public void manufacturerIsNull() {
        Car car = new Car(null, "DD-AB-123", 4);

        Set<ConstraintViolation<Car>> constraintViolations =
            validator.validate(car);

        assertEquals(1, constraintViolations.size());
        assertEquals("may not be null", constraintViolations.iterator().next().getMessage());
    }

    @Test
    public void licensePlateTooShort() {
        Car car = new Car("Morris", "D", 4);

        Set<ConstraintViolation<Car>> constraintViolations = 
            validator.validate(car);

        assertEquals(1, constraintViolations.size());
        assertEquals("size must be between 2 and 14", constraintViolations.iterator().next().getMessage());
    }
    
    @Test
    public void seatCountTooLow() {
        Car car = new Car("Morris", "DD-AB-123", 1);

        Set<ConstraintViolation<Car>> constraintViolations =
            validator.validate(car);

        assertEquals(1, constraintViolations.size());
        assertEquals("must be greater than or equal to 2", constraintViolations.iterator().next().getMessage());
    }

    @Test
    public void carIsValid() {
        Car car = new Car("Morris", "DD-AB-123", 2);

        Set<ConstraintViolation<Car>> constraintViolations =
            validator.validate(car);

        assertEquals(0, constraintViolations.size());
    }
}

In the setUp() method we get a Validator instance from the ValidatorFactory. A Validator instance is thread-safe and may be reused multiple times. For this reason we store it as field of our test class. We can use the Validator now to validate the different car instances in the test methods.

The validate() method returns a set of ConstraintViolation instances, which we can iterate in order to see which validation errors occurred. The first three test methods show some expected constraint violations:

  • The @NotNull constraint on manufacturer is violated in manufacturerIsNull()

  • The @Size constraint on licensePlate is violated in licensePlateTooShort()

  • The @Min constraint on seatCount is violated in seatCountTooLow()

If the object validates successfully, validate() returns an empty set.

Note that we only use classes from the package javax.validation from the Bean Validation API. As we don't reference any classes of the RI directly, it would be no problem to switch to another implementation of the API, should that need arise.

In this chapter we will see in more detail how to use Hibernate Validator to validate constraints for a given entity model. We will also learn which default constraints the Bean Validation specification provides and which additional constraints are only provided by Hibernate Validator. Let's start with how to add constraints to an entity.

Constraints in Bean Validation are expressed via Java annotations. In this section we show how to annotate an object model with these annotations. We have to differentiate between three different type of constraint annotations - field-, property-, and class-level annotations.

When validating an object that implements an interface or extends another class, all constraint annotations on the implemented interface and parent class apply in the same manner as the constraints specified on the validated object itself. To make things clearer let's have a look at the following example:


Our well-known class Car from ??? is now extended by RentalCar with the additional property rentalStation. If an instance of RentalCar is validated, not only the @NotNull constraint on rentalStation is validated, but also the constraint on manufacturer from the parent class.

The same would hold true, if Car were an interface implemented by RentalCar.

Constraint annotations are aggregated if methods are overridden. If RentalCar would override the getManufacturer() method from Car any constraints annotated at the overriding method would be evaluated in addition to the @NotNull constraint from the super-class.

The Bean Validation API does not only allow to validate single class instances but also complete object graphs. To do so, just annotate a field or property representing a reference to another object with @Valid. If the parent object is validated, all referenced objects annotated with @Valid will be validated as well (as will be their children etc.). See Example 2.6, “Adding a driver to the car”.



If an instance of Car is validated, the referenced Person object will be validated as well, as the driver field is annotated with @Valid. Therefore the validation of a Car will fail if the name field of the referenced Person instance is null.

Object graph validation also works for collection-typed fields. That means any attributes that are

  • arrays

  • implement java.lang.Iterable (especially Collection, List and Set)

  • implement java.util.Map

can be annotated with @Valid, which will cause each contained element to be validated, when the parent object is validated.


If a Car instance is validated, a ConstraintValidation will be created, if any of the Person objects contained in the passengers list has a null name.

Note

null values are getting ignored when validating object graphs.

The Validator interface is the main entry point to Bean Validation. In Section 5.1, “Configuration and ValidatorFactory” we will first show how to obtain an Validator instance. Afterwards we will learn how to use the different methods of the Validator interface.

The Validator interface contains three methods that can be used to either validate entire entities or just a single properties of the entity.

All three methods return a Set<ConstraintViolation>. The set is empty, if the validation succeeds. Otherwise a ConstraintViolation instance is added for each violated constraint.

All the validation methods have a var-args parameter which can be used to specify, which validation groups shall be considered when performing the validation. If the parameter is not specified the default validation group (javax.validation.Default) will be used. We will go into more detail on the topic of validation groups in Section 2.3, “Validating groups”

As we will see in Chapter 3, Creating custom constraints each constraint definition must define a default message descriptor. This message can be overridden at declaration time using the message attribute of the constraint. You can see this in Example 2.13, “Driver”. This message descriptors get interpolated when a constraint validation fails using the configured MessageInterpolator. The interpolator will try to resolve any message parameters, meaning string literals enclosed in braces. In order to resolve these parameters Hibernate Validator's default MessageInterpolator first recursively resolves parameters against a custom ResourceBundle called ValidationMessages.properties at the root of the classpath (It is up to you to create this file). If no further replacements are possible against the custom bundle the default ResourceBundle under /org/hibernate/validator/ValidationMessages.properties gets evaluated. If a replacement occurs against the default bundle the algorithm looks again at the custom bundle (and so on). Once no further replacements against these two resource bundles are possible remaining parameters are getting resolved against the attributes of the constraint to be validated.

Since the braces { and } have special meaning in the messages they need to be escaped if they are used literally. The following The following rules apply:

  • \{ is considered as the literal {

  • \} is considered as the literal }

  • \\ is considered as the literal \

If the default message interpolator does not fit your requirements it is possible to plug a custom MessageInterpolator when the ValidatorFactory gets created. This can be seen in Chapter 5, Bootstrapping.

Groups allow you to restrict the set of constraints applied during validation. This makes for example wizard like validation possible where in each step only a specified subset of constraints get validated. The groups targeted are passed as var-args parameters to validate, validateProperty and validateValue. Let's have a look at an extended Car with Driver example. First we have the class Person (Example 2.12, “Person”) which has a @NotNull constraint on name. Since no group is specified for this annotation its default group is javax.validation.Default.

Note

When more than one group is requested, the order in which the groups are evaluated is not deterministic. If no group is specified the default group javax.validation.Default is assumed.


Next we have the class Driver (
Example 2.13, “Driver”) extending Person. Here we are adding the properties age and hasDrivingLicense. In order to drive you must be at least 18 (@Min(18)) and you must have a driving license (@AssertTrue). Both constraints defined on these properties belong to the group DriverChecks. As you can see in Example 2.14, “Group interfaces” the group DriverChecks is just a simple tagging interface. Using interfaces makes the usage of groups type safe and allows for easy refactoring. It also means that groups can inherit from each other via class inheritance.

Note

The Bean Validation specification does not enforce that groups have to be interfaces. Non interface classes could be used as well, but we recommend to stick to interfaces.



Last but not least we add the property passedVehicleInspection to the Car class (
Example 2.15, “Car”) indicating whether a car passed the road worthy tests.


Overall three different groups are used in our example. Person.name, Car.manufacturer, Car.licensePlate and Car.seatCount all belong to the Default group. Driver.age and Driver.hasDrivingLicense belong to DriverChecks and last but not least Car.passedVehicleInspection belongs to the group CarChecks.
Example 2.16, “Drive away” shows how passing different group combinations to the Validator.validate method result in different validation results.

Example 2.16. Drive away

public class GroupTest {

    private static Validator validator;

    @BeforeClass
    public static void setUp() {
        ValidatorFactory factory = Validation.buildDefaultValidatorFactory();
        validator = factory.getValidator();
    }

    @Test
    public void driveAway() {
        // create a car and check that everything is ok with it.
        Car car = new Car( "Morris", "DD-AB-123", 2 );
        Set<ConstraintViolation<Car>> constraintViolations = validator.validate( car );
        assertEquals( 0, constraintViolations.size() );

        // but has it passed the vehicle inspection?
        constraintViolations = validator.validate( car, CarChecks.class );
        assertEquals( 1, constraintViolations.size() );
        assertEquals("The car has to pass the vehicle inspection first", constraintViolations.iterator().next().getMessage());

        // let's go to the vehicle inspection
        car.setPassedVehicleInspection( true );
        assertEquals( 0, validator.validate( car ).size() );

        // now let's add a driver. He is 18, but has not passed the driving test yet
        Driver john = new Driver( "John Doe" );
        john.setAge( 18 );
        car.setDriver( john );
        constraintViolations = validator.validate( car, DriverChecks.class );
        assertEquals( 1, constraintViolations.size() );
        assertEquals( "You first have to pass the driving test", constraintViolations.iterator().next().getMessage() );

        // ok, John passes the test
        john.passedDrivingTest( true );
        assertEquals( 0, validator.validate( car, DriverChecks.class ).size() );

        // just checking that everything is in order now
        assertEquals( 0, validator.validate( car, Default.class, CarChecks.class, DriverChecks.class ).size() );
    }
}


First we create a car and validate it using no explicit group. There are no validation errors, even though the property passedVehicleInspection is per default false. However, the constraint defined on this property does not belong to the default group.

Next we just validate the CarChecks group which will fail until we make sure that the car passes the vehicle inspection.

When we then add a driver to the car and validate against DriverChecks we get again a constraint violation due to the fact that the driver has not yet passed the driving test. Only after setting passedDrivingTest to true the validation against DriverChecks will pass.

Last but not least, we show that all constraints are passing by validating against all defined groups.

By default, constraints are evaluated in no particular order and this regardless of which groups they belong to. In some situations, however, it is useful to control the order of the constraints evaluation. In our example from Section 2.3, “Validating groups” we could for example require that first all default car constraints are passing before we check the road worthiness of the car. Finally before we drive away we check the actual driver constraints. In order to implement such an order one would define a new interface and annotate it with @GroupSequence defining the order in which the groups have to be validated.

Note

If at least one constraints fails in a sequenced group none of the constraints of the follwoing groups in the sequence get validated.


The usage of the new sequence could then look like in Example 2.18, “Usage of a group sequence”.


The @GroupSequence annotation also fulfills a second purpose. It allows you to redefine what the Default group means for a given class. To redefine Default for a class, place a @GroupSequence annotation on the class. The defined groups in the annotation express the sequence of groups that substitute Default for this class. Example 2.19, “RentalCar” introduces a new class RentalCar with a redfined default group. With this definition the check for all three groups can be rewritten as seen in Example 2.20, “testOrderedChecksWithRedefinedDefault”.



Note

Due to the fact that there cannot be a cyclic dependency in the group and group sequence definitions one cannot just add Default to the sequence redefining Default for a class. Instead the class itself should be added!

Hibernate Validator implements all of the default constraints specified in Bean Validation as well as some custom ones. Table 2.2, “Built-in constraints” list all constraints available in Hibernate Validator.

Table 2.2. Built-in constraints

AnnotationPart of Bean Validation SpecificationApply onUseHibernate Metadata impact
@AssertFalseyesfield/propertycheck that the annotated element is false.none
@AssertTrueyesfield/propertycheck that the annotated element is true.none
@DecimalMaxyesfield/property. Supported types are BigDecimal, BigInteger, String, byte, short, int, long and the respective wrappers of the primitive types.The annotated element must be a number whose value must be lower or equal to the specified maximum. The parameter value is the string representation of the max value according to the BigDecimal string representation.none
@DecimalMinyesfield/property. Supported types are BigDecimal, BigInteger, String, byte, short, int, long and the respective wrappers of the primitive types.The annotated element must be a number whose value must be higher or equal to the specified minimum. The parameter value is the string representation of the min value according to the BigDecimal string representation.none
@Digits(integer=, fraction=)yesfield/property. Supported types are BigDecimal, BigInteger, String, byte, short, int, long and the respective wrappers of the primitive types.Check whether the property is a number having up to integer digits and fraction fractional digits.Define column precision and scale.
@Emailnofield/property. Needs to be a string.Check whether the specified string is a valid email address.none
@Futureyesfield/property. Supported types are java.util.Date and java.util.Calendar.Checks whether the annotated date is in the future.none
@Length(min=, max=)nofield/property. Needs to be a string.Validate that the annotated string is between min and max included.none
@Maxyesfield/property. Supported types are BigDecimal, BigInteger, String, byte, short, int, long and the respective wrappers of the primitive types.Checks whether the annotated value is less than or equal to the specified maximum.Add a check constraint on the column.
@Minyesfield/property. Supported types are BigDecimal, BigInteger, String, byte, short, int, long and the respective wrappers of the primitive types.Check whether the annotated value is higher than or equal to the specified minimum.Add a check constraint on the column.
@NotNullyesfield/propertyCheck that the annotated value is not null.Column(s) are not null.
@NotEmptynofield/property. Needs to be a string.Check if the string is not null nor empty.none
@Nullyesfield/propertyCheck that the annotated value is null.none
@Pastyesfield/property. Supported types are java.util.Date and java.util.Calendar.Checks whether the annotated date is in the past.none
@Pattern(regex=, flag=)yesfield/property. Needs to be a string.Check if the annotated string match the regular expression regex.none
@Range(min=, max=)nofield/property. Supported types are BigDecimal, BigInteger, String, byte, short, int, long and the respective wrappers of the primitive types.Check whether the annotated value lies between (inclusive) the specified minimum and maximum.none
@Size(min=, max=)yesfield/property. Supported types are String, Collection, Map and arrays.Check if the annotated element size is between min and max (inclusive).Column length will be set to max.
@Validyesfield/propertyPerform validation recursively on the associated object.none

Note

On top of the parameters indicated in Table 2.2, “Built-in constraints” each constraint supports the parameters message, groups and payload. This is a requirement of the Bean Validation specification.

In some cases these built-in constraints will not fulfill your requirements. In this case you can literally in a minute write your own constraints. We will discuss this in Chapter 3, Creating custom constraints

Though the Bean Validation API defines a whole set of standard constraint annotations one can easily think of situations in which these standard annotations won't suffice. For these cases you are able to create custom constraints tailored to your specific validation requirements in a simple manner.

To create a custom constraint, the following three steps are required:

Let's write a constraint annotation, that can be used to express that a given string shall either be upper case or lower case. We'll apply it later on to the licensePlate field of the Car class from Chapter 1, Getting started to ensure, that the field is always an upper-case string.

First we need a way to express the two case modes. We might use String constants, but a better way to go is to use a Java 5 enum for that purpose:


Now we can define the actual constraint annotation. If you've never designed an annotation before, this may look a bit scary, but actually it's not that hard:


An annotation type is defined using the @interface keyword. All attributes of an annotation type are declared in a method-like manner. The specification of the Bean Validation API demands, that any constraint annotation defines

Besides those three mandatory attributes (messge, groups and payload) we add another one allowing for the required case mode to be specified. The name value is a special one, which can be omitted upon using the annotation, if it is the only attribute specified, as e.g. in @CheckCase(CaseMode.UPPER).

In addition we annotate the annotation type with a couple of so-called meta annotations:

  • @Target({ METHOD, FIELD, ANNOTATION_TYPE }): Says, that methods, fields and annotation declarations may be annotated with @CheckCase (but not type declarations e.g.)

  • @Retention(RUNTIME): Specifies, that annotations of this type will be available at runtime by the means of reflection

  • @Constraint(validatedBy = CheckCaseValidator.class): Specifies the validator to be used to validate elements annotated with @CheckCase

  • @Documented: Says, that the use of @CheckCase will be contained in the JavaDoc of elements annotated with it

Next, we need to implement a constraint validator, that's able to validate elements with a @CheckCase annotation. To do so, we implement the interface ConstraintValidator as shown below:


The ConstraintValidator interface defines two type parameters, which we set in our implementation. The first one specifies the annotation type to be validated (in our example CheckCase), the second one the type of elements, which the validator can handle (here String).

In case a constraint annotation is allowed at elements of different types, a ConstraintValidator for each allowed type has to be implemented and registered at the constraint annotation as shown above.

The implementation of the validator is straightforward. The initialize() method gives us access to the attribute values of the annotation to be validated. In the example we store the CaseMode in a field of the validator for further usage.

In the isValid() method we implement the logic, that determines, whether a String is valid according to a given @CheckCase annotation or not. This decision depends on the case mode retrieved in initialize(). As the Bean Validation specification recommends, we consider null values as being valid. If null is not a valid value for an element, it should be annotated with @NotNull explicitely.

The passed-in ConstraintValidatorContext could be used to raise any custom validation errors, but as we are fine with the default behavior, we can ignore that parameter for now.

Now that our first custom constraint is completed, we can use it in the Car class from the Chapter 1, Getting started chapter to specify that the licensePlate field shall only contain upper-case strings:


Finally let's demonstrate in a little test that the @CheckCase constraint is properly validated:


Looking at the licensePlate field of the Car class in Example 3.5, “Applying the CheckCase constraint”, we see three constraint annotations already. In complexer scenarios, where even more constraints could be applied to one element, this might become a bit confusing easily. Furthermore, if we had a licensePlate field in another class, we would have to copy all constraint declarations to the other class as well, violating the DRY principle.

This problem can be tackled using compound constraints. In the following we create a new constraint annotation @ValidLicensePlate, that comprises the constraints @NotNull, @Size and @CheckCase:


To do so, we just have to annotate the constraint declaration with its comprising constraints (btw. that's exactly why we allowed annotation types as target for the @CheckCase annotation). As no additional validation is required for the @ValidLicensePlate annotation itself, we don't declare a validator within the @Constraint meta annotation.

Using the new compound constraint at the licensePlate field now is fully equivalent to the previous version, where we declared the three constraints directly at the field itself:


The set of ConstraintViolations retrieved when validating a Car instance will contain an entry for each violated composing constraint of the @ValidLicensePlate constraint. If you rather prefer a single ConstraintViolation in case any of the composing constraints is violated, the @ReportAsSingleViolation meta constraint can be used as follows:


The key to enable XML configuration for Hibernate Validator is the file validation.xml. If this file exists in the classpath its configuration will be applied when the ValidationFactory gets created. Example 4.1, “validation-configuration-1.0.xsd” shows a model view of the xsd valiation.xml has to adhere to.


Example 4.2, “validation.xml” shows the several configuration options of validation.xml.


Warning

There can only be one validation.xml in the classpath. If more than one is found an exception is thrown.

All settings shown in the validation.xml are optional and in the case of Example 4.2, “validation.xml” show the defaults used within Hibernate Validator. The node default-provider allows to choose the Bean Validation provider. This is useful if there is more than one provider in the classpath. message-interpolator, traversable-resolver and constraint-validator-factory allow to customize the javax.validation.MessageInterpolator, javax.validation.TraversableResolver resp. javax.validation.ConstraintValidatorFactory. The same configuration options are also available programmatically through the javax.validation.Configuration. In fact XML configuration will be overriden by values explicitly specified via the API. It is even possible to ignore the XML configuration completely via Configuration.ignoreXmlConfiguration(). See also Chapter 5, Bootstrapping.

Via the constraint-mapping you can list an arbitrary number of additional XML files containing the actual constraint configuration. See Section 4.2, “Mapping constraints”.

Last but not least, you can specify provider specific properties via the property nodes. Hibernate Validator does currently not make use of any custom properties.

Expressing constraints in XML is possible via files adhering to the xsd seen in Example 4.3, “validation-mapping-1.0.xsd”. Note that these mapping files are only processed if listed via constraint-mapping in your validation.xml.


Example 4.4, “constraints-car.xml” shows how our classes Car and RentalCar from Example 2.15, “Car” resp. Example 2.19, “RentalCar” could be mapped in XML.

Example 4.4. constraints-car.xml

<constraint-mappings xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
                     xsi:schemaLocation="http://jboss.org/xml/ns/javax/validation/mapping validation-mapping-1.0.xsd"
                     xmlns="http://jboss.org/xml/ns/javax/validation/mapping">
    <default-package>org.hibernate.validator.quickstart</default-package>
    <bean class="Car" ignore-annotations="true">
        <field name="manufacturer">
            <constraint annotation="javax.validation.constraints.NotNull"/>
        </field>
        <field name="licensePlate">
            <constraint annotation="javax.validation.constraints.NotNull"/>
        </field>
        <field name="seatCount">
            <constraint annotation="javax.validation.constraints.Min">
                <element name="value">2</element>
            </constraint>
        </field>
        <field name="driver">
            <valid/>
        </field>
        <getter name="passedVehicleInspection" ignore-annotations="true">
            <constraint annotation="javax.validation.constraints.AssertTrue">
                <message>The car has to pass the vehicle inspection first</message>
                <groups>
                    <value>CarChecks</value>
                </groups>
                <element name="max">10</element>
            </constraint>
        </getter>
    </bean>
    <bean class="RentalCar" ignore-annotations="true">
        <class ignore-annotations="true">
            <group-sequence>
                <value>RentalCar</value>
                <value>CarChecks</value>
            </group-sequence>
        </class>
    </bean>
    <constraint-definition annotation="org.mycompany.CheckCase" include-existing-validator="false">
        <validated-by include-existing-validators="false">
            <value>org.mycompany.CheckCaseValidator</value>
        </validated-by>
    </constraint-definition>
</constraint-mappings>

The XML configuration is closely mirroring the programmatic API. For this reason it should suffice to just add some comments. default-package is used for all fields where a classname is expected. If the specified class is not fully qualified the configured default package will be used. Every mapping file can then have several bean nodes, each describing the constraints on the entity with the specified class name.

Warning

A given entity can only be configured once across all configuration files. If the same class is configured more than once an exception is thrown.

Settings ignore-annotations to true means that constraint annotations placed on the configured bean are ignored. The default for this value is true. ignore-annotations is also available for the nodes class, fields and getter. If not explicitly specified on these levels the configured bean value applies. Otherwise do the nodes class, fields and getter determine on which level the constraints are placed (see Section 2.1, “Defining constraints”). The constraint node is then used to add a constraint on the corresponding level. Each constraint definition must define the class via the annotation attribute. The constraint attributes required by the Bean Validation specification (message, groups and payload) have dedicated nodes. All other constraint specific attributes are configured using the the element node.

The class node also allows to reconfigure the default group sequence (see Section 2.3.2, “Redefining the default group sequence of a class”) via the group-sequence node.

Last but not least, the list of ConstraintValidators associated to a given constraint can be altered via the constraint-definition node. The annotation attribute represents the constraint annotation being altered. The validated-by elements represent the (ordered) list of ConstraintValidator implementations associated to the constraint. If include-existing-validator is set to false, validators defined on the constraint annotation are ignored. If set to true, the list of ConstraintValidators described in XML are concatenated to the list of validators described on the annotation.

We already seen in Section 5.1, “Configuration and ValidatorFactory” the easiest way to create a Validator instance - Validation.buildDefaultValidatorFactory. In this chapter we have a look at the other methods in javax.validation.Validation and how they allow to configure several aspects of Bean Validation at bootstrapping time.

The different bootstrapping options allwow, amongst other things, to bootstrap any Bean Validation implementation on the classpath. Generally, an available provider is discovered by the Java Service Provider mechanism. A Bean Validation implementation includes the file javax.validation.spi.ValidationProvider in META-INF/services. This file contains the fully qualified classname of the ValidationProvider of the implementation. In the case of Hibernate Validator this is org.hibernate.validator.HibernateValidator.

Note

If there are more than one Bean Validation implementation providers in the classpath and Validation.buildDefaultValidatorFactory() is used, there is no guarantee which provider will be chosen. To enforce the provider Validation.byProvider() should be used.

There are three different methods in the Validation class to create a Validator instance. The easiest in shown in Example 5.1, “Validation.buildDefaultValidatorFactory()”.


You can also use the method Validation.byDefaultProvider() which will allow you to configure several aspects of the created Validator instance:


We will learn more about MessageInterpolator, TraversableResolver and ConstraintValidatorFactory in the following sections.

Last but not least you can ask for a Configuration object of a specific Bean Validation provider. This is useful if you have more than one Bean Validation provider in your classpath. In this situation you can make an explicit choice about which implementation to use. In the case of Hibernate Validator the Validator creation looks like:


The usage of the TraversableResolver has so far not been discussed. The idea is that in some cases, the state of a property should not be accessed. The most obvious example for that is a lazy loaded property or association of a Java Persistence provider. Validating this lazy property or association would mean that its state would have to be accessed triggering a load from the database. Bean Validation controls which property can and cannot be accessed via the TraversableResolver interface (see Example 5.7, “TraversableResolver interface”).

Example 5.7. TraversableResolver interface

/**
 * Contract determining if a property can be accessed by the Bean Validation provider
 * This contract is called for each property that is being either validated or cascaded.
 *
 * A traversable resolver implementation must be thread-safe.
 *
 */
public interface TraversableResolver {
    /**
     * Determine if the Bean Validation provider is allowed to reach the property state
     *
     * @param traversableObject object hosting <code>traversableProperty</code> or null  
     *                          if validateValue is called
     * @param traversableProperty the traversable property.
     * @param rootBeanType type of the root object passed to the Validator.
     * @param pathToTraversableObject path from the root object to
     *        <code>traversableObject</code>
     *        (using the path specification defined by Bean Validator).
     * @param elementType either <code>FIELD</code> or <code>METHOD</code>.
     *
     * @return <code>true</code> if the Bean Validation provider is allowed to
     *         reach the property state, <code>false</code> otherwise.
     */
     boolean isReachable(Object traversableObject,
                         Path.Node traversableProperty,
                         Class<?> rootBeanType,
                         Path pathToTraversableObject,
                         ElementType elementType);

    /**
     * Determine if the Bean Validation provider is allowed to cascade validation on
     * the bean instance returned by the property value
     * marked as <code>@Valid</code>.
     * Note that this method is called only if isReachable returns true for the same set of
     * arguments and if the property is marked as <code>@Valid</code>
     *
     * @param traversableObject object hosting <code>traversableProperty</code> or null
     *                          if validateValue is called
     * @param traversableProperty the traversable property.
     * @param rootBeanType type of the root object passed to the Validator.
     * @param pathToTraversableObject path from the root object to
     *        <code>traversableObject</code>
     *        (using the path specification defined by Bean Validator).
     * @param elementType either <code>FIELD</code> or <code>METHOD</code>.
     *
     * @return <code>true</code> if the Bean Validation provider is allowed to
     *         cascade validation, <code>false</code> otherwise.
     */
     boolean isCascadable(Object traversableObject,
                          Path.Node traversableProperty,
                          Class<?> rootBeanType,
                          Path pathToTraversableObject,
                          ElementType elementType);
}


Hibernate Validator provides two TraversableResolvers out of the box which will be enabled automatically depending on your environment. The first is the DefaultTraversableResolver which will always return true for isReachable() and isTraversable(). The second is the JPATraversableResolver which gets enabled when Hibernate Validator gets used in combination with JPA 2. In case you have to provide your own resolver you can do so again using the Configuration object as seen in
Example 5.8, “Providing a custom TraversableResolver”.


Last but not least, there is one more configuration option to discuss, the ConstraintValidatorFactory. The default ConstraintValidatorFactory provided by Hibernate Validator requires a public no-arg constructor to instantiate ConstraintValidator instances (see Section 3.1.2, “The constraint validator”). Using a custom ConstraintValidatorFactory offers for example the possibility to use dependency injection in constraint implementations. The configuration of the custom factory is once more via the Configuration (Example 5.9, “Providing a custom ConstraintValidatorFactory”).


The interface you have to implement is:


Hibernate Validator is intended to be used to implement multi-layered data validation, where constraints are expressed in a single place (the annotated domain model) and checked in various different layers of the application.

Hibernate Validator integrates with both Hibernate and all pure Java Persistence providers.

Hibernate Validator has a built-in Hibernate event listener - org.hibernate.cfg.beanvalidation.BeanValidationEventListener - which is part of Hibernate Annotations (as of Hibernate 3.5.x). Whenever a PreInsertEvent, PreUpdateEvent or PreDeleteEvent occurs, the listener will verify all constraints of the entity instance and throw an exception if any constraint is violated. Per default objects will be checked before any inserts or updates are made by Hibernate. Pre deletion events will per default not trigger a validation. You can configure the groups to be validated per event type using the properties javax.persistence.validation.group.pre-persist, javax.persistence.validation.group.pre-update and javax.persistence.validation.group.pre-remove. The values of these properties are the comma seperated, fully specified class names of the groups to validate. Example 6.1, “Manual configuration of BeanValidationEvenListener” shows the default values for these properties. In this case they could also be omitted.

On constraint violation, the event will raise a runtime ConstraintViolationException which contains a set of ConstraintViolations describing each failure.

If Hibernate Validator is present in the classpath, Hibernate Annotations (or Hibernate EntityManager) will use it transparently. To avoid validation even though Hibernate Validator is in the classpath set javax.persistence.validation.mode to none.

Note

If the beans are not annotated with validation annotations, there is no runtime performance cost.

In case you need to manually set the event listeners for Hibernate Core, use the following configuration in hibernate.cfg.xml:


If you are using JPA 2 and Hibernate Validator is in the classpath the JPA2 specification requires that Bean Validation gets enabled. The properties javax.persistence.validation.group.pre-persist, javax.persistence.validation.group.pre-update and javax.persistence.validation.group.pre-remove as described in Section 6.2.1, “Hibernate event-based validation” can in this case be configured in persistence.xml. persistence.xml also defines a node validation-mode while can be set to AUTO, CALLBACK, NONE. The default is AUTO.

In a JPA 1 you will have to create and register Hibernate Validator yourself. In case you are using Hibernate EntityManager you can add a customized version of the BeanValidationEventListener described in Section 6.2.1, “Hibernate event-based validation” to your project and register it manually.

Last but not least, a few pointers to further information. A great source for examples is the Bean Validation TCK which can is available for anonymous access in the Hibernate SVN repository. Alternatively you can view the tests using Hibernate's fisheye installation. The JSR 303 specification itself is also a great way to deepen your understanding of Bean Validation resp. Hibernate Validator.

If you have any furhter questions to Hibernate Validator or want to share some of your use cases have a look at the Hibernate Validator Wiki and the Hibernate Validator Forum.

In case you would like to report a bug use Hibernate's Jira instance. Feedback is always welcome!