XPath usage for System.XML namespace objects

[Plater]

I'm having a bit of trouble getting XPath expressions to work to select nodes.

I am working with the google maps results.
A shortened example xml:

<?xml version="1.0" encoding="UTF-8" ?> 
 
<kml xmlns="http://earth.google.com/kml/2.0">
    <Document>
        <Placemark>
            <name>Route</name> 
            <description>
                <![CDATA[ Distance: 2,164 mi (about 1 day 7 hours)<br/>Map data ©2009 Sanborn, Tele Atlas  ]]> 
          </description>
      </Placemark>
  </Document>
</kml>
 

Now really there is a lot more to that xml, but what I am concerned with is the <Placemark> node that has the name of "Route", and then I want the contents of the description node.

After populating an XmlDocument doc, I tried to use this:
doc.SelectSingleNode("/kml/Document/placemark[name='route']/description");
and
doc.SelectNodes("/kml/Document/placemark[name='route']/description");

They always turn up no results.
I even tried:
doc.SelectNodes("//description");
which also produced no results.

I was able to get it by using a non-xpath expression:
doc.ChildNodes[1].ChildNodes[0].LastChild["description"];

But seems very unstable as it relies on those indexes existing.

Does anyone know what is wrong with the XPath I used? Is it a namespace issue (I set up no namespaces)?

[Frinavale]

Hey Plater, I moved this to the XML forum since the experts here know how to use XPath expressions.

[Plater]

I can use XPath in other systems (like in FF's javascript) just fine, everything works as expected. The problem was with how .NET uses it.
I'll leave it here for now.

[jkmyoung]

Namespace. xmlns="http://earth.google.com/kml/2.0"
Try looking at XMLNamespaceManager when using SelectNodes.
http://msdn.microsoft.com/en-us/library/4bektfx9.aspx

Don't forget you'll have to prefix each node. So, assuming ge is the namespace,
doc.SelectNodes("/ge:kml/ge:Document/ge:placemark[ge:name='route']/ge:description");

[Plater]

Drat I was hoping to not have to deal with that.
I never should have left string functions.
hehe everytime I try to use XML to perform a task, I bail on it and decide its faster to just string function.

[jkmyoung]

True, String functions would be faster if you don't have any other description nodes. Also, I forgot you'd probably have to change case, eg "route" != "Route"