Java: Using XPath to loop over nodes and extract specific subnode values

Question

Hi

I understand from various googling that it makes more sense to try to extract your data from XML using XPath than using DOM looping. At the moment I have implemented my solution to use DOM, but the code is verbose, it feels untidy and unmaintainable, so I would like to switch to a cleaner XPath solution.

Lets say I have this structure:

<products>
    <product>
        <title>Some title 1</title>
        <image>Some image 1</image>
    </product>
    <product>
        <title>Some title 2</title>
        <image>Some image 2</image>
    </product>
    ...
</products>

I want to be able to run a for loop for each of the <product> elements, and inside this for loop, extract the title and image node values.

My java looks like:

InputStream is = conn.getInputStream();          
DocumentBuilder builder = DocumentBuilderFactory.newInstance().newDocumentBuilder();
Document doc = builder.parse(is);
XPathFactory factory = XPathFactory.newInstance();
XPath xpath = factory.newXPath();
XPathExpression expr = xpath.compile("/products/product");
Object result = expr.evaluate(doc, XPathConstants.NODESET);
NodeList products = (NodeList) result;
for (int i = 0; i < products.getLength(); i++) {
    Node n = products.item(i);
    if (n != null && n.getNodeType() == Node.ELEMENT_NODE) {
        Element product = (Element) n;
        // do some DOM navigation to get the title and image
    }
}

So inside my for loop I get each <product> as a Node, which is cast to an Element.

Can I simply use my instance of XPathExpression to compile and run another XPath on the Node or the Element?

Any help much appreciated.

Answer 1

Yes, you can always do like this -

XPathFactory factory = XPathFactory.newInstance();
XPath xpath = factory.newXPath();
XPathExpression expr = xpath.compile("/products/product");
Object result = expr.evaluate(doc, XPathConstants.NODESET);
expr = xpath.compile("title"); // The new xpath expression to find 'title' within 'product'.

NodeList products = (NodeList) result;
for (int i = 0; i < products.getLength(); i++) {
    Node n = products.item(i);
    if (n != null && n.getNodeType() == Node.ELEMENT_NODE) {
        Element product = (Element) n;
        NodeList nodes = (NodeList)  expr.evaluate(product,XPathConstants.NODESET); //Find the 'title' in the 'product'
        System.out.println("TITLE: " + nodes.item(0).getTextContent()); // And here is the title 
    }
}    

Here I have given example of extracting the 'title' value. In same way you can do for 'image'

Answer 2

I'm not a big fan of this approach because you have to build a document (which might be expensive) before you can apply XPaths to it.

I've found VTD-XML a lot more efficient when it comes to applying XPaths to documents, because you don't need to load the whole document into memory. Here is some sample code:

final VTDGen vg = new VTDGen();
vg.parseFile("file.xml", false);
final VTDNav vn = vg.getNav();
final AutoPilot ap = new AutoPilot(vn);

ap.selectXPath("/products/product");
while (ap.evalXPath() != -1) {
    System.out.println("PRODUCT:");

    // you could either apply another xpath or simply get the first child
    if (vn.toElement(VTDNav.FIRST_CHILD, "title")) {
        int val = vn.getText();
        if (val != -1) {
            System.out.println("Title: " + vn.toNormalizedString(val));
        }
        vn.toElement(VTDNav.PARENT);
    }
    if (vn.toElement(VTDNav.FIRST_CHILD, "image")) {
        int val = vn.getText();
        if (val != -1) {
            System.out.println("Image: " + vn.toNormalizedString(val));
        }
        vn.toElement(VTDNav.PARENT);
    }
}

Also see this post on Faster XPaths with VTD-XML.