How to ignore (efficiently) most part of an XML structure when unmarshalling with JAXB?

Question

When processing a web service response with a quite complex XML structure, I am only interested with a very small subset of information. Lets consider the using JAXB has to be used in this context.

As an example, lets say I am only interested in retrieving d (which can be modeled as a single JAXB bean):

What is the fastest recommended way to ignore everything else but retrieve d?

I'm guessing that you can't use xpath, that you have to use jaxb, correct? — Hovercraft Full Of Eels, Sep 20 '11 at 21:43
Hi, Thanks for this question. actually, I am also trying to do something similar but for some reason, it's not working as expected. I tried the things mentioned in this answer as well. I tried this approach and the example from the authors blog but still the JAXB Marshalling does not work as expected with `@XmlPath`. Seems like the XML produced with and without `@XmlPath` are the same. Can you please once look into this example and provide your answer: https://stackoverflow.com/questions/67500565/xmlpath-has-no-impact-during-the-jaxb-marshalling — BATMAN_2008, May 12 '21 at 10:38

score 3 · Answer 1 · answered Sep 20 '11 at 21:54

I think the fastest way would depend on exactly how the xml is formatted. E.g. you could create your own InputStream that wraps the real InputStream and just be on the lookout for "<d>" to trigger you to start passing data through to jaxb.

e.g.

   InputStream is = // get real stream
   JAXBContext jc = JAXBContext.newInstance( "com.acme.foo" );
   Unmarshaller u = jc.createUnmarshaller();
   Object o = u.unmarshal( new MySpecialStream(is) );

If you needed a more xml-like approach then you could create your own XMLStreamReader that only passed on certain calls if you were inside a d element.

e.g.

   JAXBContext jc = JAXBContext.newInstance( "com.acme.foo" );
   Unmarshaller u = jc.createUnmarshaller();
   XMLStreamReader xsr = XMLInputFactory.newInstance().createXMLStreamReader( ... );
   Object o = u.unmarshal( new MyXSRWrapper(xsr) );

Your StAX approach can be implemented using a `StreamFilter` - http://stackoverflow.com/questions/7492128/how-to-ignore-efficiently-most-part-of-an-xml-structure-when-unmarshalling-with/7493327#7493327 — bdoughan, Sep 21 '11 at 00:56

score 3 · Accepted Answer · edited May 23 '17 at 09:59

Note: I'm the EclipseLink JAXB (MOXy) lead and a member of the JAXB 2 (JSR-222) expert group.

There are a couple of different ways that you could handle this use case:

Option #1 - StreamFilter Any JAXB Implementation

You could use a StAX XMLStreamReader with a StreamFilter to filter out the portions of the XML document that you don't wish to map to:

JAXB filtered parsing

Option #2 - @XmlPath in MOXy JAXB

You could use the @XmlPath extension in MOXy:

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

    @XmlPath("b2/d/text()")
    private String d;

}

For More Information

score 0 · Answer 3 · answered Sep 20 '11 at 22:05

0

Use XPath to select the desired element as a DOM node and unmarshal using that selected node.

answered Sep 20 '11 at 22:05

forty-two

12,204
2
26
36

Sometimes it's a massive waste of time and space creating a DOM tree just for this. – biziclop Sep 20 '11 at 22:09
@biziclop So is premature optimization. What does 'fastest' refer to, development time or processing time, and what is more expensive? – forty-two Sep 20 '11 at 22:17
I don't know, but as I pointed out, DOM is often simply not an option. – biziclop Sep 21 '11 at 13:31

How to ignore (efficiently) most part of an XML structure when unmarshalling with JAXB?

3 Answers3