Efficient way to construct an XML document from a string containing well formed XML for navigation?

Question

I have a string that contains well formed xml in it. I want to navigate the XML in that string to extract the text in certain nodes. How do I efficiently accomplish this using a built-in .NET class. Which .NET XML class would you use and why?

Many thanks for your help.

Note 1: Linq is not available to me. Note 2: Editing the XML is not important. Read-only access is what I need.

Answer 1

I would use XmlDocument.Load() to get a DOM from the string. Then you can traverse it using the appropriate DOM methods or XPATH as needed.

Answer 2

It depends on the structure of XML. If it is relatively simple, then the most efficient way to wrap the string into StringReader, and then wrap that into XmlReader. The benefit is that you won't have to create an XML tree in memory, copying data from the string - you'll just read nodes one by one.

If the document structure is complicated enough, you might need (or want) a DOM - in which case XDocument.Parse should do the trick.

Answer 3

For navigation? Probably XPathDocument:

string s = @"<xml/>";
XPathDocument doc = new XPathDocument(new StringReader(s));

From MSDN,

Provides a fast, read-only, in-memory representation of an XML document by using the XPath data model.

Unlike XmlDocument etc, it is optimised for readonly usage; more efficient but less powerful (i.e. you can't edit it). For notes on how to query it, see here.

Answer 4

For speed, use an XmlReader:

using (StringReader sr = new StringReader(myString))
using (XmlReader xr = XmlReader.Create(sr))
{
    while (xr.Read())
    {
        if (xr.NodeType == XmlNodeType.Element && xr.Name == "foo")
        {
            Console.WriteLine(xr.ReadString());
        }
    }
}

The above prints out the text content of every element named "foo" in the XML document. (Well, sort of. ReadString doesn't handle nested elements very gracefully.)

Using an XPathDocument is slower, because the entire document gets parsed before you can start searching it, but it has the merit of simplicity:

using (StringReader sr = new StringReader(myString))
{
    XPathDocument d = new XPathDocument(sr);
    foreach (XPathNavigator n in d.CreateNavigator().Select("//foo/text()"))
    {
        Console.WriteLine(n.Value);
    }
}

If you're not concerned with performance or memory utilization, it's simplest to use an XmlDocument:

XmlDocument d = new XmlDocument();
d.LoadXml(myString);
foreach (XmlNode n in d.SelectNodes("//foo/text()"))
{
   Console.WriteLine(n.Value);
}