Reading XML

[Gregg Bolinger]

Ok, this may be a stupid question, but guess what, I don't care.
From my limited knowledge of the various API's for reading/writing XML files in Java it seems overly complicated in the fact that I have to write my own special classes to read and write an XML file. So I have a lot of custom parsing and what not if I am using my own XML files that I created.
Now, it could be that I just need to supply my XML API with the DTD of the XML file and all of a sudden things get easier, but I don't know. That's kind of what I'm asking.
Basically, in my make believe ideal world, an API to read/write XML files should be as simple as providing the API my XML file, maybe a DTD then I can just read/write whatever object I want. For example

<user role="admin">
  <username>gdboling</username>
  <password>bolinger</password>
</user>

And to read this might look something like

Parser foo = Parser.parse("users.xml");
System.out.println(foo.getNode("user").getNode("username").toString());
System.out.println(foo.getNode("user").getAttribute("role").toString());

And the output of that would of course be

gdboling
admin

Maybe this is possible and I am just missing something. Any hints, tips, suggestions, or simply telling me I'm an idiot will suffice.
Thanks.

[Lasse Koskela]

The standard JAXP API includes classes for parsing XML documents into a DOM document tree, which you can then interrogate for details a bit like in your make-believe example.
It goes like this:

// First you get a factory object
DocumentBuilderFactory factory = DocumentBuilderFactory.newInstance();
// Then you configure the factory as you wish
factory.setValidating(true); // we want the factory to create validating parsers
// Then we ask for a parser (called DocumentBuilder, don't ask me why)
DocumentBuilder builder = factory.newDocumentBuilder();
// Then you ask the parser to parse the XML document
Document document = builder.parse(new FileInputStream("/tmp/user.xml"));
// And finally you traverse the DOM tree and pick those parts you need
// The org.w3c.dom.* API is a bit complex but you'll manage... 
Element user = document.getDocumentElement(); // this returns the "root" element
// attributes are easy to get
String role = user.getAttribute("role");
// this is one of the methods for selecting a NodeList of child elements
Element usernameElem = user.getElementsByNodeName("username").item(0);
// you need to first get the Text node that's a child of <username>
// and you ask for *that* node's value, not the enclosing element's
String username = usernameElem.getFirstChild().getNodeValue();
// the same thing for <password>
Element passwordElem = user.getElementsByNodeName("password").item(0);
String password = passwordElem.getFirstChild().getNodeValue();

[Lasse Koskela]

Originally posted by Gregg Bolinger:
...or simply telling me I'm an idiot will suffice.

Damn. I almost missed this.
You're an idiot

[Lasse Koskela]

One more thing. Check out JavaAlmanac.com and search for "dom" or "xml".

[Gregg Bolinger]

Element usernameElem = user.getElementsByNodeName("username").item(0);
Why am I getting item(0)? What is item(0)?

[Gregg Bolinger]

Ok, I might be starting to get the hang of this. So the value between a specific element is the NodeValue, <username>gdboling</username>, gdboling is the NodeValue and the NodeName is the name of the element which would be username in that case.
I'm still not clear as to why I have to get item(0). In what case might I get say item(1)? I realize I probably don't have an item(1) in my user.xml file but what would I add so that it does?

[Lasse Koskela]

Originally posted by Gregg Bolinger:
Ok, I might be starting to get the hang of this. So the value between a specific element is the NodeValue, <username>gdboling</username>, gdboling is the NodeValue and the NodeName is the name of the element which would be username in that case.

In DOM, everything is a node (org.w3c.dom.Node). In your "user" document, each of the elements are nodes (of type org.w3c.dom.Element) and each piece of plain text is a node (or type org.w3c.dom.Text). The javadocs for org.w3c.dom.Node explains what is the NodeValue for each type of node. For Elements it's null, for Text nodes it's the text. Note that even though <username>gdboling</username> would intuitively mean that the "username" element has one child node of type Text with value "gdboling", it might have two child nodes: one could be a Text node with value "gdb" and the other with value "oling".

Originally posted by Gregg Bolinger:
I'm still not clear as to why I have to get item(0). In what case might I get say item(1)? I realize I probably don't have an item(1) in my user.xml file but what would I add so that it does?

The item(int) method needs to be called because the getElementsByName(String) methods return a list of nodes that match the given name. In my example, I knew that the only child with the name "username" is the username I want so I just took the first (0-based index) elements off the list.