Fork me on GitHub

Tutorial 15

E15 Reading a complex plist structure and deriving the XPath from method names

In this tutorial we address quite a few challenges in accessing the data in plist files. XML plists (XML Property Lists) files are key/value based data storages frequently used by OS X applications. The challenge is that the data is not stored in the key element, but besides. A key element is followed by an additional element holding the data. The name of the data containing element depends on the type of data. So we need some real XPath magic to cover this :) The second challenge is that we don't like to have XPath expressions specified for all data in the file. Instead we like to derive the key value of the expression from the method name. Last but not least we show how to switch the DTD validation of, in case your not on an OS X system.

XML Data

<?xml version="1.0" encoding="UTF-8"?>
<plist version="1.0">
    <string>Charles Robert Darwin</string>
      <string>A man who dares to waste one hour of time has not discovered the value of life.</string>
      <string>I have tried lately to read Shakespeare, and found it so intolerably dull that it nauseated me.</string>
      <string>I love fools' experiments. I am always making them.</string>


public interface PList {
    String getAuthor();
    String[] getLines();
    int getBirthdate();

Example Code

public class TestPlistAccess extends TutorialTestCase {
     * Every plist file has a DTD referenced which might be unavailable if you are on a non MacOs system.
     * So we tweak the XBProjector configuration to ignore the DTD.
    private final class NonValidatingXMLFactoriesConfig extends DefaultXMLFactoriesConfig {
        public DocumentBuilderFactory createDocumentBuilderFactory()  {
            try {
            final DocumentBuilderFactory factory = super.createDocumentBuilderFactory();
            factory.setFeature("", false);
            return factory;
            catch (final ParserConfigurationException e) {
                throw new RuntimeException(e);
     * The projection in this tutorial does not define any XPath expression to the data.
     * Instead we like to derive the XPath from the method name. Although this could be done
     * in a single XPath expression, we split it up for two cases for readability:
     * - Selecting all children of the following element of the key element, if an array is expected
     * and
     * - Selecting the first following element of the key element, in all other cases.
    public class PListExternalizer extends ExternalizerAdapter {
        public String resolveXPath(final String annotationValue, final Method method, final Object[] args) {
            final String keyValue = method.getName().substring(3);
            if (method.getReturnType().isArray()) {
                return  "/plist/dict/key[.=\""+keyValue+"\"]/following-sibling::*[1]/child::*";
           return  "/plist/dict/key[.=\""+keyValue+"\"]/following-sibling::*[1]";
    public void testReadPList() throws IOException {
        final XBProjector projector = new XBProjector(new NonValidatingXMLFactoriesConfig());
        projector.config().setExternalizer(new PListExternalizer());
        final PList plist =;
        System.out.println(plist.getAuthor()+" ("+plist.getBirthdate()+")");
        for (final String line:plist.getLines()) {