Question

Ok. I have an attribute in an xml document that I know will occur more than once. Using C# I loop through all the nodes that have this attribute. I know how to count the occurrence of an element using xpath…

count("//x/y@b")

and so on.
But is there a way that I can get the n-th value of a node that I am on… for example

<?xml version="1.0"?>
<x>
   <y/>
   <y/>
   <y/>
</x>

Let’s say I was looping through that programatically using c#. And lets say I was on the second element. Is there any way using xpath that I could figure out that I am on the 2nd node? I guess I am just trying to find my position in the iteration. Any ideas? Currently scouring the internet. If I find it out I will be sure to let you know.

Thanks.

UPDATE: CAN’T SEEM to get my stuff to work
Ok. I thought I would update my question. I can’t seem to get any of your suggestions working…

<Template>
 <TemplateData>
  <ACOData>
   <POPULATION_PATIENT_ID>6161</POPULATION_PATIENT_ID>
   <PATIENT_ID>4329</PATIENT_ID>
  </ACOData>
  <ACOData>
   <POPULATION_PATIENT_ID>5561</POPULATION_PATIENT_ID>
   <PATIENT_ID>4327</PATIENT_ID>
  </ACOData>
  <ACOData>
   <POPULATION_PATIENT_ID>6160</POPULATION_PATIENT_ID>
   <PATIENT_ID>4321</PATIENT_ID>
  </ACOData>
  <ACOData>
   <POPULATION_PATIENT_ID>5561</POPULATION_PATIENT_ID>
   <PATIENT_ID>4320</PATIENT_ID>
  </ACOData>

That is the XML that I am using. But I can’t seem to get the correct count. I am always coming up with zero?

encounter = Int32.Parse((patElm.CreateNavigator().Evaluate("count(/Template/TemplateData/ACOData/POPULATION_PATIENT_ID[.='" + populationPatID + "']/preceding-sibling::ACOData/POPULATION_PATIENT_ID[.='"+populationPatID+"'])")).ToString());

The above is the code that I am attempting to use to get the correct value… Note my count function

count(/Template/TemplateData/ACOData/POPULATION_PATIENT_ID[.='" + populationPatID + "']/preceding-sibling::ACOData/POPULATION_PATIENT_ID[.='"+populationPatID+"'])"

Answer 1

To get the second such element in the document use:

(//x/y[@b])[2]

Suppose you want to go the other way. That is, you have one of these nodes and you want to know its overall position. In general, for any expression <expr> the following is true:

$n = count((<expr>)[$n]/preceding::*[count(.|<expr>)=count(<expr>)])

That is, the position of the Nth element selected by <expr> can be found by counting all the preceding elements also selected by that expression. Using similar techniques, we can find the position of some node that would be selected by a more general expression, within the set of all nodes selected by that expression.

For example, suppose we have the following document:

<x>
    <y b="true"/>
    <y b="true"/>
    <y/>
    <y/>
    <x><y b="true"/><y/><y b="true">77</y></x>
    <y/>
    <y/>
</x>

And we want to know the position in the document of the node at /*/*/y[.='77'] among all nodes selected by //x/y[@b]. Then use the following expression:

count(/*/*/y[.='77']/preceding::*[count(.|//x/y[@b])=count(//x/y[@b])]) + 1

A more specific one-off solution looks like this:

count(/*/*/y[.='77']/preceding::y[parent::x and @b]) + 1

Result (in both cases):

4

Note: It’s assumed that /*/*/y[.='77'] and (<expr>)[$n] above actually select some node in the document. If not, the result will be an erroneous 1 due to adding 1 to the result of the count. For this reason, this method is probably most useful when working on a context node or when it is guaranteed that your initial expression selects a node. (Of course, initial error checking can be employed, as well.)