How to retrieve node after node of XML tree using Xpath?

Question

First, I must say that I find Xpath as a very nice parser , and I guess pretty powerful when comparing it to other parsers .

Given the following code :

  DocumentBuilderFactory domFactory = 
  DocumentBuilderFactory.newInstance();
  domFactory.setNamespaceAware(true); 
  DocumentBuilder builder = domFactory.newDocumentBuilder();
  Document doc = builder.parse("input.xml");
  XPath xpath = XPathFactory.newInstance().newXPath();

If I wanted to find the first node of Round 1 & Door 1 , here :

<Game>
    <Round>
        <roundNumber>1</roundNumber>
        <Door>
            <doorName>abd11</doorName>
            <Value>
                <xVal1>0</xVal1>
                <xVal2>25</xVal2>
                <pVal>0.31</pVal>
            </Value>
            <Value>
                <xVal1>25</xVal1>
                <xVal2>50</xVal2>
                <pVal>0.04</pVal>
            </Value>
            <Value>
                <xVal1>50</xVal1>
                <xVal2>75</xVal2>
                <pVal>0.19</pVal>
            </Value>
            <Value>
                <xVal1>75</xVal1>
                <xVal2>100</xVal2>
                <pVal>0.46</pVal>
            </Value>
        </Door>
        <Door>
            <doorName>vvv1133</doorName>
            <Value>
                <xVal1>60</xVal1>
                <xVal2>62</xVal2>
                <pVal>1.0</pVal>
            </Value>
        </Door>
    </Round>
    <Round>
        <roundNumber>2</roundNumber>
        <Door>
            <doorName>eee</doorName>
            <Value>
                <xVal1>0</xVal1>
                <xVal2>-25</xVal2>
                <pVal>0.31</pVal>
            </Value>
            <Value>
                <xVal1>-25</xVal1>
                <xVal2>-50</xVal2>
                <pVal>0.04</pVal>
            </Value>
            <Value>
                <xVal1>-50</xVal1>
                <xVal2>-75</xVal2>
                <pVal>0.19</pVal>
            </Value>
            <Value>
                <xVal1>-75</xVal1>
                <xVal2>-100</xVal2>
                <pVal>0.46</pVal>
            </Value>
        </Door>
        <Door>
            <doorName>cc</doorName>
            <Value>
                <xVal1>-60</xVal1>
                <xVal2>-62</xVal2>
                <pVal>0.3</pVal>
            </Value>
            <Value>
                <xVal1>-70</xVal1>
                <xVal2>-78</xVal2>
                <pVal>0.7</pVal>
            </Value>
        </Door>
    </Round>
</Game>

I'll do this :

 XPathExpression expr = xpath.compile("//Round[1]/Door[1]/Value[1]/*/text()");      
  Object result = expr.evaluate(doc, XPathConstants.NODESET);
  NodeList nodes = (NodeList) result;

and if I wanted the second node of Round 1 & Door 1 then :

XPathExpression expr = xpath.compile("//Round[1]/Door[1]/Value[2]/*/text()");  

but how do I do this using a loop , since I don't know how much Value-nodes I have , meaning how can I do this using a loop , where each iteration I retrieve 3 (I mean the xVal1 , xVal2 and pVal values ) more values of a Value node !?

The reasons for asking for this are :

1. I don't know how much Round-s I have

2. I don't know how much Value-s I have

3. I don't want to declare every time a new XPathExpression

Thanks .

Solution

Option 1 - Iterate over all Value elements in the document. Only one evaluation required, but difficult to know which Round or Door element the Value belongs to.

NodeList result = (NodeList) xpath.evaluate("//Round/Door/Value/*/text()", doc, XPathConstants.NODESET);

Option 2 - Iterate over each Round, Door and Value elements separately. Requires more evaluations but the context is easily known. If index is required, it is easy to add a counter to the loops.

// Get all rounds and iterate over them
NodeList rounds = (NodeList) xpath.evaluate("//Round", doc, XPathConstants.NODESET);
for (Node round : rounds) {
  // Get all doors and iterate over them
  NodeList doors = (NodeList) xpath.evaluate("Door", round, XPathConstants.NODESET);
  for (Node door : doors) {
    // Get all values and iterate over them
    NodeList values = (NodeList) xpath.evaluate("Value/*/text()", door, XPathConstants.NODESET);
    for (Node value : values) {
      // Do something
    }
  }
}

Option 3 - Do some combination of the above depending on your requirements

Note that I've removed the expression compilation step to shorten the example. It should be re-added to improve performance.

Comments

[Dimitre Novatchev]

Use the XPath count() function to find the number of nodes, then use a (nested) loop:

[Ollie]

Why not several expressinos? Isn't that what you want (e.g. first get list of Rounds, then Doors, then Values?)

[JAN]

@erikxiv : The problem is that each Round has 2 fields , one is "roundNumber" and the second is "Door" . Now with a single query , i.e. XPathExpression , I can only get either but not both ? because from what I see at the moment , I have to do 2 queries - one for the field "roundNumber" and another for the "Door" of each Round ... and so on

[JAN]

@DimitreNovatchev: I've tried count but it doesn't work , take a look : if I write this : "double count = (Double) xpath.evaluate("count(//Round/Door[1]/Value)", doc, XPathConstants.NUMBER);" then I get "count == 8" ......... if I write instead "int k = 1; double count = (Double) xpath.evaluate("count(//Round/Door[k]/Value)", doc, XPathConstants.NUMBER);" , then I get "count == 0"

[Dimitre Novatchev]

@ron: This is one of the most common XPath FAQ: use: count((//Round/Door)[1]/Value) and then count((//Round/Door)[$k]/Value)

[majorbanzai]

Which NodeList class are you using that implements Iterable?

[Ollie]

@majorbanzai Are you thinking about the enhanced for-loops syntax? I suspect you are right in that the code sample won't work as NodeList is not an Iterable. Does make the example more readable though...

[viruskimera]

Nodelist cannot be iterated like that.