Hi,

This has come up several times before, but I cannot persuade the DOM parser to ignore whitespace in an XML document. I've read through all the previous queries on the problem, and I

1. validate against a schema (this works, as it complains if I change the XML)
2. use xerces 2.5, which ought to validate against schemas correctly
3. set setIgnoringElementContentWhitespace(true)
4. in the schema explicitly set mixed="false" in every complexType definition

I've modified the DomEcho02 example from the tutorial, and it still shows me text nodes around every child node. Has anybody got any ideas why this is not working the way it should?

The code setting up the factory is:

      if (argv.length != 2) {
        System.err.println("Usage: java DomEcho filename schema");
        System.exit(1);
      }
 
      DocumentBuilderFactory factory =
        DocumentBuilderFactory.newInstance();
      factory.setValidating(true);
      factory.setNamespaceAware(true);
      factory.setIgnoringComments(true);
      factory.setIgnoringElementContentWhitespace(true);
      factory.setAttribute(
        "http://java.sun.com/xml/jaxp/properties/schemaLanguage",
        "http://www.w3.org/2001/XMLSchema");
      try {
        factory.setAttribute(
          "http://java.sun.com/xml/jaxp/properties/schemaSource",
          new InputSource(new FileInputStream(argv[1])));
        DocumentBuilder builder = factory.newDocumentBuilder();
        document = builder.parse(new File(argv[0]));
        makeFrame();
 
      } ....

I'd appreciate any help on this one!

Adriaan

 

 

http://forum.java.sun.com/thread.jsp?forum=34&thread=430252

 

 

Just an update on what I did in the end: I've written a small recursive function that traverses the tree and throws out all empty text nodes. Not a nice solution at all - especially considering the inefficiencies of building a DOM tree and then throwing half of it away again.

From the documentation I still conclude that the white space text nodes should not be generated - after all, the schema does not allow them. So my conclusion is that this simply does not work in xerces yet. Have to try again with the next version...

 

 

This has been a frustrating issue for me as well, and it has been no joy to wade through lots of comments of the kind "just skip past the blank text nodes".

Here is how I got it to work in JDK 5.0 RC.

1) Get a DOMImplementation the non-portable way:

DocumentBuilderFactory factory = DocumentBuilderFactory.newInstance();
DocumentBuilder builder = factory.newDocumentBuilder();
DOMImplementation domImpl = builder.getDOMImplementation();

(If you call

DOMImplementationRegistry registry = DOMImplementationRegistry.newInstance();
DOMImplementation domImpl = registry.getDOMImplementation("XML 3.0");

then you lose in step 2--the standard domImpl can't get the "LS" feature--how lame is that?)

2) Make a LSParser

DOMImplementationLS ls = (DOMImplementationLS) domImpl.getFeature("LS", "3.0");
LSInput in = ls.createLSInput();
in.setByteStream(new FileInputStream(filename));
LSParser parser = ls.createLSParser(DOMImplementationLS.MODE_SYNCHRONOUS, "http://www.w3.org/2001/XMLSchema");

3) That's it--parse, and it just works. No more white space!

Document doc = parser.parse(in);

It is completely incomprehensible to me why the JAXP parser can't do this correctly. Amusingly enough, the bug http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=4867706, reported in May 2003, is marked as a "request for enhancement".

Cheers,

Cay