在XML文件中测试XPath的最佳方法是什么？

Question

Lately I've been using XPathDocument and XNavigator to parse an XML file for a given XPath and attribute. It's been working very well, when I know in advance what the XPath is.

Sometimes though, the XPath will be one of several possible XPath values, and I'd like to be able to test whether or not a given XPath exists.

In case I'm getting the nomenclature wrong, here's what I'm calling an XPath - given this XML blob:

<foo>
    <bar baz="This is the value of the attribute named baz">
</foo>

I might be looking for what I'm calling an XPath of "//foo/bar" and then reading the attribute "baz" to get the value.

Example of the code that I use to do this:

XPathDocument document = new XPathDocument(filename);
XPathNavigator navigator = document.CreateNavigator();
XPathNavigator node = navigator.SelectSingleNode("//foo/bar");
if(node.HasAttributes)
{
    Console.WriteLine(node.GetAttribute("baz", string.Empty));
}

Now, if the call to navigator.SelectSingleNode fails, it will return a NullReferenceException or an XPathException. I can catch both of those and refactor the above into a test to see whether or not a given XPath returns an exception, but I was wondering whether there was a better way?

I didn't see anything obvious in the Intellisense. XPathNavigator has .HasAttributes and .HasChildren but short of iterating through the path one node at a time, I don't see anything nicer to use.

Answer 1

If you've given valid XPath but it doesn't match anything, SelectSingleNode won't throw a NullReferenceException - it will just return null.

If you pass SelectSingleNode some syntactically invalid XPath, that's when it will throw an XPathException .

So normally, you'd just need to test whether the returned value was null or not.

Answer 2

 var baz = navigator.SelectSingleNode("//foo/bar/@baz");
 if (baz != null) Console.WriteLine(baz);

Answer 3

I think it is not good to create an XMLNode object by executing navigator.SelectSingleNode(...).

You have to use navigator.Evaluate() instead:

if (Convert.ToBoolean(navigator.Evaluate(@"boolean(//foo/bar)"))) {...}

Answer 4

From memory, may contain errors.

XDocument doc = XDocument.Load("foo.xml");

var att = from a in doc.Descendants("bar")
          select a.Attribute("baz")

foreach (var item in att) {
    if (item != null) { ... }
}

Answer 5

If node == null then node.HasAttributes will throw a NullReferenceException . This situation will occur when //foo/bar does not match any elements in the XML document.

Answer 6

var node = XDocument.Load(filename)
                    .Descendants("bar")
                    .SingleOrDefault(e=>e.Attribute("baz") != null);

if (node != null) Console.WriteLine(node.Attribute("baz").Value);

Answer 7

I would probably be more specific in my xpath.

        var doc = XDocument.Load(fileName);

        var results = from r in doc.XPathSelectElements("/foo/bar[count(@baz) > 0]")
                      select r.Attribute("baz");

        foreach (String s in results)
            Console.WriteLine(s);