Skip to main content

Notice: This Wiki is now read only and edits are no longer possible. Please see: for the plan.

Jump to: navigation, search

UserGuide/JPA/Using the Canonical Model Generator (ELUG)

Canonical Model Generation

With the new Criteria API from the JPA 2.0 spec, EclipseLink can now produce static

metamodel classes that correspond to the entities, mapped superclasses, and embeddable classes in the persistence unit. A static metamodel class models the persistent state and relationships of the corresponding managed class.

For portability, EclipseLink generates this canonical metamodel as defined in section of the specification. EclipseLink generates the static metamodel classes by using the the annotation processor tool (APT) in conjunction with its existing metadata processing of annotations.

For more information on APT, see:

Metadata Processing

The annotation processor tool directly ties into the regular deployment metadata processing. Which means before generating the metamodel classes, persistence units are processed as they normally would be according to the spec giving you the following behavior:

  • Persistence unit property processing
  • XML mapping file merging and override
  • Full annotation and xml processing of Entities, Embeddables and MappedSuperclasses.
  • Full support of EclipseLink extensions, including annotations and eclipelink-orm.xml

Configuring and using within Eclipse Galileo

Use the following procedure to configure the metamodel generation within the Eclipse IDE.

  1. Select your project in the Package Explorer and select Project > Properties. The Properties dialog appears.
  2. Select Java Compiler and ensure you are using JDK 1.6 (or higher).
  3. Expand the Java Compiler element and select Annotation Processing.
  4. In the Annotation Processing area, enable the Enable annotation processing option.
  5. In the Generated source directory field, enter the directory in which to generate the metamodel classes.
  6. By default, Processor options are not needed. However, EclipseLink provides you with custom generation options if you would like to configure you metamodel classes differently then defined in the spec. See the EclipseLink custom processor options section below for supported processor options.
  7. Expand the Annotation Processing element and select Factory Path.
  8. Click Add External JARs and add the following libraries (JARs). Refer to the the 1.2.0 RC4 Library names for details.
    • The EclipseLink JAR (eclipselik.jar)
    • The Java persistence 2.0 preview JAR
    • A JAR that contains the enabling services file that specifies the name of the annotation processor tool.

Service Jar

Note: The annotation processor does not log to the Console window. Instead, messages are logged to the Error Log window. To display this window, select Window > Show View > Error Log from the Eclipse menu.

Configuring and using within ANT/javac

Use the following procedure to configure the metamodel generation within ANT/javac:

  1. Add modelgen jar to the javac classpath.
  2. Use the -processor option on javac command line.
  3. In the following examples, options are specified with using -A (for example, javac -Aeclipselink.persistencexml=META-INF/sub/persistence.xml)

ANT example

    <target name="model-compile" description="compile models that need canonical processing">
            deprecation="${javac.deprecation}" failonerror="true">
            <include name="${MODEL_PATH}/**/*.java"/>
            <compilerarg value="-Aeclipselink.persistencexml=${PERSISTENCE_XML_PATH}" compiler="javac1.6"/>
            <classpath refid="${COMPILE_PATH_REF}"/>

EclipseLink custom processor options

The following table describes the EclipseLink custom options:

Option Description
eclipselink.persistencexml The full resource name in which to look for the persistence XML files. If not specified the default META-INF/persistence.xml will be used.
eclipselink.persistenceunits A comma deliminated list of persistence unit names that will be used when generating the canonical model. By default all persistence units available in all persistence XML files will be used.
Note: When specifying multiple persistence units you cannot include a persistence unit with a comma ( , ) in its name.
eclipselink.canonicalmodel.prefix The prefix that will be added to the start of the class name of any generated canonical model class. By default the prefix is not used.
eclipselink.canonicalmodel.suffix The suffix that will be added to the end of the class name of any generated canonical model class. The default suffix value is _ (an underscore). If specified, this property value must be a non-empty string that contains valid characters for use in a Java class name.
eclipselink.canonicalmodel.subpackage A sub-package name that can be used to have the canonical model generator generate its classes in a sub-package of the package where the corresponding entity class is located. By default the canonical model classes are generated into the same package as the entity classes.

EclipseLink 1.2.0 RC4 Library names

Currently the target release of EclipseLink 1.2.0 is RC4 which uses the following jar files names in the ZIP installer. Please note that the ZIP installer is the only distribution of EclipseLink including the modelgen JAR file.

  • Services file (enables the processor): Service Jar
  • javax.persistence library:
  • EclipseLink jar:


After following the configuration steps above if no classes are generated, check the following:

Do you have a persistence XML file?

Verify that you have a Persistence XML file. By default, the file is located in META-INF/persistence.xml.

If you name your persistence XML differently, or place it in a different directory, you must use the eclipselink.persistencexml processor option. See EclipseLink Custom Processor Options for more information.

Are you using an extended EclipseLink ORM mapping file?

If you use an extended EclipseLink ORM mapping file that is not listed in your Persistence XML file, EclipseLink will not automatically discover the mapping file (similar to the orm.xml file). You must explicitly specify the mapping file:

Note: Future versions of EclipseLink will include the ability to automatically discover the eclipselink-orm.xml file for the metamodel class generation. In a regular deployment scenario the eclipselink-orm discovery still applies.

Are XML changes not being reflected in the generated model classes?

After making an XML change within Eclipse, the the generated model classes are not updated to reflect the change. Unlike changes to the model class, XML changes are not automatically reflected, since the annotation processor is not aware of these changes.

To reflect XML changes in your generated metamodel classes, clean the project. Select Project > Clean... from the Eclipse menu.

If you have made extensive XML changes (including the removal of mapping files or deletion of entities), you must restart the annotation processor tool:

  1. Select Project > Properties from the Eclipse menu.
  2. Expand Java Compiler element and select Annotation Processing.
  3. In the Annotation Processing area of the Properties window, deselect the Enable annotation processing option.
  4. Click OK. A window appears, indicating that you have changed the annotation processing settings and Eclipse must rebuild the project.
  5. Click Yes.
  6. After Eclipse rebuilds the project, return to the Annotations Processing area of the Properties page, and enable the Enable annotation processing option.
  7. Click OK. Again, a window appears, indicating that you have changed the annotation processing settings and Eclipse must rebuild the project.
  8. Click Yes. Eclipse rebuilds the project. The generated model classes now reflect your latest XML changes.

Copyright Statement

Back to the top