DTD's and getElementById

Question

Hi folks...
Im getting myself in too deep here ... I wish to build a DOM object containing NodeElements I add ... I need to find these elements by getElementById .. and I keep getting null exceptions when using that method.
I know that the DOM recommendation says you have to define ID's on elements ... and I tried to use a DTD Ive made myself.
..but I cant figure out how !!
This is the code I have experimented with:
class XML {
     public static void main(String[] args) {
          try {
               DocumentBuilderFactory factory = DocumentBuilderFactory.newInstance();
               DocumentBuilder b = factory.newDocumentBuilder();
               factory.setNamespaceAware(true);
               factory.setValidating(true);
               DOMImplementation dom = b.getDOMImplementation();
               DocumentType doctype = dom.createDocumentType("root", "test", "file:///test.dtd");
               System.out.println(doctype.getNotations().getLength());
               Document doc = dom.createDocument("", "root", doctype);
               Element elem2 = doc.createElement("person");
               Attr id = doc.createAttribute("ID");
               id.setValue("test");
               elem2.setAttributeNode(id);
               doc.appendChild(elem);
               doc.getDocumentElement().appendChild(elem2);
               System.out.println(doc.getElementById("test"));
          } catch(Exception e) {
               System.out.println("error: " + e.getMessage() + " " + e.getStackTrace() + " " + e.getCause());
..the doctype doesn't apply to the document ... why? .. Ive tried without the file:// .. notation too - and getElementById keeps returning null..
Anybody got a answer ??? Ive looked through numerous questions in this Forum...I just still don't get it
tnx in advance.

Answer 1

True but remarkably unhelpful. I did a grep across the JDK docs and the associated W3 docs last night; there doesn't seem to be any mechanism for defining the ID. Perhaps they didn't implement it.
This can be implemented by the user... just traverse through the tree (assuming using DOM) until you find an element with attribute "ID" matching the value you want, or you could build a B-tree using SAX or DOM to the same end. But what a pain, and hardly optimized the way the library function presumably should be.