Using XPath - Java returns the NodeList in document order - how can I fix it ?

[david ron]

Hello ,

I need to fix a known example is XPath 1.0 - the reversed axis .


Given the following XML file as an example for what is written afterwards :

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE stock SYSTEM "new.dtd">
<stock>
    <book num="myBook1">
        <title>Lindsy Boxer</title>
        <author>James Patterson</author>
        <publisher>LittleBig</publisher>
        <price>18.21</price>
        <chapter>
            <title>Alex Cross Is Back - Chapter A</title>
            <paragraph>
                This is the <emph>first</emph> paragraph.
                <image file="alexCrossImage.gif"/>
                afetr image...
            </paragraph>
            <paragraph>
                This is the <emph>second</emph> paragraph.
                <image file="alexCrossImageAnother.gif"/>
                afetr image...
            </paragraph>
        </chapter>
        <chapter>
            <title>Along Came A Spider - Chapter B</title>
            <section>
                <title>Along Came A Spider - Chapter B - section 1</title>
                <paragraph>
                    This is the <emph>first</emph>paragraph for chapter TWO section ONE.
                    <image file="Chapter_B_firstParagraphImage.gif"/>
                    afetr image...
                </paragraph>
                <paragraph>
                    This is the <emph>second</emph> paragraph for chapter TWO section ONE.
                    <image file="Chapter_B_secondParagraphImage.gif"/>
                    afetr image...
                </paragraph>
            </section>
        </chapter>
        <chapter>
            <title>Chapter C</title>
            <paragraph>
                This chapter has no images and only one paragraph
            </paragraph>
        </chapter>
    </book>
    <book num="myBook2">
        <title>Jack Reacher Series</title>
        <author>Lee Child</author>
        <author>Jonny White</author>
        <publisher>Pocket</publisher>
        <price>5.99</price>
        <chapter>
            <title>Jack Reacher - Chapter ONE</title>
        </chapter>
        <chapter>
            <title>Jack Reacher - Chapter TWO</title>
            <paragraph>
                This is the <emph>second</emph> paragraph of SECOND book chapter TWO.
                <image file="Jack_Reacher_Picture_Top_Sniper_US_Army.gif"/>
                afetr image...
            </paragraph>
        </chapter>
    </book>
    <book num="myBook3">
        <title>Alex Cross - Double Cross</title>
        <author>James Patterson</author>
        <publisher>Spectra</publisher>
        <price>17.30</price>
        <chapter>
            <title>Alex Cross - Double Cross - Chapter A</title>
        </chapter>
    </book>
</stock>

I need to fix a known problem in XPath 1.0 , which known as the Reverse Axis . A little example might clarify :
When I enter the following query :

Object myQuery= xpath.evaluate("stock/book/chapter[3]/preceding-sibling::chapter//title",doc,XPathConstants.NODESET);

Output would be :

<title>Alex Cross Is Back - Chapter A</title>
<title>Along Came A Spider - Chapter B</title>
<title>Along Came A Spider - Chapter B - section 1</title>

However what I want is when my XPath uses the preceding-sibling axis ,I want to return the nodes in a reverse order , i.e. :

      <title>Along Came A Spider - Chapter B</title>
      <title>Along Came A Spider - Chapter B - section 1</title>
      <title>Alex Cross Is Back - Chapter A</title>

Apparently, Java returns the NodeList in document order.

As you can see ,Java returns the nodes by the order of the nodes in the XML file , but what I need is , when XPath

uses the preceding-sibling axis ,then return the nodes in a reverse order .

This is not the one and only command where this kind of "behavior" is happening (returning mixed nodes) .

My objective is to fix this , I've tried to change XPath's evaluate , but it's still doesn't work .

Can anyone give some directions / hints how to do this ?

I've started with rewriting the evaluate method , but nothing so far .

I'm not allowed to use XPath 2.0.

I used the following code to produce the example above :

       DocumentBuilderFactory domFactory = DocumentBuilderFactory.newInstance();
        domFactory.setNamespaceAware(true); // never forget this!
        DocumentBuilder builder = domFactory.newDocumentBuilder();
        Document doc = builder.parse("new.xml");
        //create an XPath factory
        XPathFactory factory = XPathFactory.newInstance();
        //create an XPath Object
        XPath xpath = factory.newXPath();
        //***********************************************************//
        Object myQuery= xpath.evaluate("stock/book/chapter[3]/preceding-sibling::chapter//title",doc,XPathConstants.NODESET);
        System.out.println("The Results are:");
        NodeList nodes = (NodeList) myQuery;
        //print the output
        for (int i = 0; i < nodes.getLength(); i++) {
            System.out.println("i: " + i);
            System.out.println("*******");
            System.out.println(nodeToString(nodes.item(i)));
            System.out.println("*******");
        }

Regards

David

[William Brogden]

Why not iterate through the NodeList from high to low?

Bill

[david ron]

William Brogden wrote:Why not iterate through the NodeList from high to low?

Bill

Try it for your self ,it still would be in incorrect order . Furthermore ,as I said ,I need to make to code work with every query , meaning
that there are queries where we'd need no reverse - hence there is no need to iterate through the NodeList from high to low - if you do that , the answer would be incorrect !!

I'd appreciate any suggestions . thanks

[William Brogden]

If you want absolute control then work with the DOM directly using the standard library instead of XPath.

Bill