JAXB & XML Infoset Preservation

In this post we will discuss how to use JAXB to preserve the XML Infoset. 


XML

The XML we will use for this example contains comments and unmapped elements.  Even though we will only map the highlighted lines, this example will demonstrate how to preserve the rest of the XML document using the JAXB Binder.

<?xml version="1.0" encoding="UTF-8"?>
<customer>
    <UNMAPPED_ELEMENT_1/>
    <name>Jane Doe</name>
    <!-- COMMENT #1 -->
    <address>
        <UNMAPPED_ELEMENT_2/>
        <street>1 A Street</street>
        <!-- COMMENT #2 -->
        <UNMAPPED_ELEMENT_3/>
        <city>Any Town</city>
    </address>
    <!-- COMMENT #3 -->
    <UNMAPPED_ELEMENT_4/>
    <phone-number type="home">555-HOME</phone-number>
    <!-- COMMENT #4 -->
    <phone-number type="cell">555-CELL</phone-number>
    <UNMAPPED_ELEMENT_5/>
    <!-- COMMENT #5 -->
</customer>

Java Model

We will use the following Java model for this example.  To save space the get/set methods have been omitted.

Customer

import java.util.List;
import javax.xml.bind.annotation.*;

@XmlRootElement
@XmlAccessorType(XmlAccessType.FIELD)
public class Customer {

    private String name;

    private Address address;

    @XmlElement(name="phone-number")
    private List<PhoneNumber> phoneNumbers;

}

Address

import javax.xml.bind.annotation.*;

@XmlAccessorType(XmlAccessType.FIELD)
public class Address {

    private String street;

    private String city;

}

Phone Number

import javax.xml.bind.annotation.*;

@XmlAccessorType(XmlAccessType.FIELD)
public class PhoneNumber {

    @XmlAttribute
    private String type;
    
    @XmlValue
    private String value;
}

Unmarshaller & Marshaller

The following code is familiar to all JAXB users.  An Unmarshaller is used to load the XML into objects, and then a Marshaller is used to save the objects back to XML.

import java.io.File;
import javax.xml.bind.*;

public class Demo {

    public static void main(String[] args) throws Exception {
        JAXBContext jc = JAXBContext.newInstance(Customer.class);

        File xml = new File("input.xml");
        Unmarshaller unmarshaller = jc.createUnmarshaller();
        Customer customer = (Customer) unmarshaller.unmarshal(xml);

        customer.getAddress().setStreet("2 NEW STREET");
        PhoneNumber workPhone = new PhoneNumber();
        workPhone.setType("work");
        workPhone.setValue("555-WORK");
        customer.getPhoneNumbers().add(workPhone);

        Marshaller marshaller = jc.createMarshaller();
        marshaller.setProperty(Marshaller.JAXB_FORMATTED_OUTPUT, true);
        marshaller.marshal(customer, System.out);
    }

}

Note that when the objects are saved back to XML using the Marshaller the unmapped content is lost:

<?xml version="1.0" encoding="UTF-8"?>
<customer>
   <name>Jane Doe</name>
   <address>
      <street>2 NEW STREET</street>
      <city>Any Town</city>
   </address>
   <phone-number type="home">555-HOME</phone-number>
   <phone-number type="cell">555-CELL</phone-number>
   <phone-number type="work">555-WORK</phone-number>
</customer>

Binder

The Binder interface was introduced in JAXB 2.0, yet many developers are not familiar with this JAXB feature. It is created from the JAXBContext, and interacts with XML in the form of a DOM. The Binder maintains an association between the Java objects and their corresponding DOM nodes.

import java.io.File;
import javax.xml.bind.*;
import javax.xml.parsers.*;
import javax.xml.transform.*;
import javax.xml.transform.dom.DOMSource;
import javax.xml.transform.stream.StreamResult;
import org.w3c.dom.*;

public class BinderDemo {

    public static void main(String[] args) throws Exception {
        DocumentBuilderFactory dbf = DocumentBuilderFactory.newInstance();
        DocumentBuilder db = dbf.newDocumentBuilder();
        File xml = new File("input.xml");
        Document document = db.parse(xml);

        JAXBContext jc = JAXBContext.newInstance(Customer.class);

        Binder<Node> binder = jc.createBinder();
        Customer customer = (Customer) binder.unmarshal(document);
        customer.getAddress().setStreet("2 NEW STREET");
        PhoneNumber workPhone = new PhoneNumber();
        workPhone.setType("work");
        workPhone.setValue("555-WORK");
        customer.getPhoneNumbers().add(workPhone);
        binder.updateXML(customer);

        TransformerFactory tf = TransformerFactory.newInstance();
        Transformer t = tf.newTransformer();
        t.transform(new DOMSource(document), new StreamResult(System.out));
    }

}

Instead of creating a new XML document, Binder applies the changes made to the object back to the original DOM.  This is how JAXB is able to preserve the XML infoset.

<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<customer>
    <UNMAPPED_ELEMENT_1/>
    <name>Jane Doe</name>
    <!-- COMMENT #1 -->
    <address>
        <UNMAPPED_ELEMENT_2/>
        <street>2 NEW STREET</street>
        <!-- COMMENT #2 -->
        <UNMAPPED_ELEMENT_3/>
        <city>Any Town</city>
    </address>
    <!-- COMMENT #3 -->
    <UNMAPPED_ELEMENT_4/>
    <phone-number type="home">555-HOME</phone-number>
    <!-- COMMENT #4 -->
    <phone-number type="cell">555-CELL</phone-number>
    <phone-number type="work">555-WORK</phone-number>
    <UNMAPPED_ELEMENT_5/>
    <!-- COMMENT #5 -->
</customer>

Note

There appears to be a bug in the Metro version of JAXB regarding this example.  You will need to ensure you use EclipseLink JAXB (MOXy).  To use the MOXy implementation you will need to add a file named jaxb.properties in with your model classes that contains the following entry:

javax.xml.bind.context.factory=org.eclipse.persistence.jaxb.JAXBContextFactory