Question

I’m trying to use XStream to import large XML files (~ 1 GB). I want to use an ObjectInputStream to read them streaming, but I need that to skip the first element and then move down a couple elements (see below, I need to import the bag_LVC:Woonplaats elements).

I’d expect that there is somehow a way to feed xstream a Reader or a InputStream which has been seeked up to my element of interest. I’ve tried to create my own HierarchicalStreamReader and move it to the right element, but you can only move up and down:

HierarchicalStreamReader xpp = new XppDriver().createReader(xmlFile);
xpp.moveDown(); // moves to <xb:antwoord>
xpp.move(); // **would move to <xb:producten>, but it private**
xpp.moveDown(); // moves to <xb:producten>
xpp.moveDown(); // moves to <product_LVC:LVC-product>
st = xstream.createObjectInputStream(xpp);
BagWoonplaats bw = (BagWoonplaats) st.readObject();

Does xstream or any of the XPP drivers have any way to let you seek to my element of interest?

xml

<xb:BAG-Extract-Deelbestand-LVC>
  <xb:antwoord>
    <xb:vraag>
    <!-- more elements here -->
    </xb:vraag>
    <xb:producten>
      <product_LVC:LVC-product>
        <bag_LVC:Woonplaats>
          <!-- elements used as fields -->
        </bag_LVC:Woonplaats>
        <bag_LVC:Woonplaats>
          <!-- elements used as fields -->
        </bag_LVC:Woonplaats>
        <!-- and so on -->

Answer 1

The solution for this problem was actually very trivial, and I’m almost embarrassed, but let’s share the answer for this problem anyway. I didn’t get the API for the HierarchicalStreamReader, after digging around in the code of XStream I managed to get an handled to it, this is how it works:

    HierarchicalStreamReader xpp = ...;
    while(xpp.hasMoreChildren()) {
        xpp.moveDown();
        // do something with the child element, and recursive into this again if appropriate
        xpp.moveUp();
    }

This is recursive method I now use to seek to the appropriate element. The location is specified by a sort of xpath expression (very basic, of course), which I have normalized in advance, into an ArrayList:

private void seekToElement(HierarchicalStreamReader xpp, ArrayList<String> xpathishArray) {
    while(xpp.hasMoreChildren()) {
        xpp.moveDown();
        if(xpp.getNodeName().equals(xpathishArray.get(0))) {
            if(xpathishArray.size() > 1) {
                xpathishArray.remove(0);
                seekToElement(xpp, xpathishArray);
            }
            return;
        }
        xpp.moveUp();
    }
}