XPath queries without namespace: XSE to the rescue
Note: this entry has moved.
Dare has posted about issues with XPath and namespaces in documents. I'm by no means encouraging this, but if you don't care about namespaces, here's a solution for you.
Using the XSE reader I proposed (and will upload real soon see the release post), you can setup a handler to catch all elements and remove any namespaces whatsoever, at reading time (i.e. long before the document is loaded), as follows:
<![CDATA[public class MyLoader
{
public XPathDocument Load(string uri)
{
TransformingXseReader xr = new TransformingXseReader(new XmlTextReader(uri));
XmlNamespaceManager mgr = new XmlNamespaceManager(xr.NameTable);
xr.AddHandler(new RelativePath().Create("*", mgr), new EventHandler(OnElement));
XPathDocument doc = new XPathDocument(xr);
return doc;
}
private void OnElement(object sender, EventArgs e)
{
TransformingXseReader tr = (TransformingXseReader) sender;
tr.ChangeName(tr.LocalName, String.Empty);
}
}]]>
The Whidbey version looks more compact:
<![CDATA[public class MyLoader
{
public XPathDocument Load(string uri)
{
TransformingXseReader xr = new TransformingXseReader(new XmlTextReader(uri));
XmlNamespaceManager mgr = new XmlNamespaceManager(xr.NameTable);
xr.AddHandler(new RelativePath().Create("*", mgr), delegate
{
xr.ChangeName(xr.LocalName, String.Empty);
}
XPathDocument doc = new XPathDocument(xr);
return doc;
}
}]]>
What you effectively get loaded is an infoset without element namespaces, therefore your XPath queries don't need to care about them anymore.
Update: read these follow-up: