Recipe 17.3 Querying the Contents of an XML Document

Problem

You have a large and complex XML document and you need to find various pieces of information, such as all the information contained within a specific element and having a particular attribute setting. You want to query the XML structure without having to iterate through all the nodes in the XML document and searching for a particular item by hand.

Solution

In order to query a database, you normally would use SQL. In order to query an XML document, you would use XPath. In .NET, this means using the System.Xml.XPath namespace and classes like XPathDocument, XPathNavigator, and XPathNodeIterator.

In the following example, we use these classes to select nodes from an XML document we construct holding members from the board game "Clue" (or "Cluedo", as it is known abroad) and their various roles. We want to be able to select the married female participants who were witnesses to the crime. In order to do this, we pass an XPath expression to query the XML data set as follows:

public static void QueryXML( )
{
    string xmlFragment = "<?xml version='1.0'?>" +
        "<Clue>" + 
        "<Participant type='Perpetrator'>Professor Plum</Participant>" +
        "<Participant type='Witness'>Colonel Mustard</Participant>" +
        "<Participant type='Witness'>Mrs. White</Participant>" +
        "<Participant type='Witness'>Mrs. Peacock</Participant>" +
        "<Participant type='Witness'>Mr. Green</Participant>" +
        "</Clue>"; 

    XmlTextReader reader = new XmlTextReader(xmlFragment, 
                            XmlNodeType.Element,null); 

    // Instantiate an XPathDocument using the XmlTextReader.  
    XPathDocument xpathDoc = new XPathDocument(reader, XmlSpace.Preserve);
    
    // get the navigator
    XPathNavigator xpathNav = xpathDoc.CreateNavigator( );
    
    // set up the query looking for the married female participants 
    // who were witnesses
    string xpathQuery = 
      "/Clue/Participant[attribute::type='Witness'][contains(text( ),'Mrs.')]";

    // get the nodeset from the query
    XPathNodeIterator xpathIter = xpathNav.Select(xpathQuery);

    // write out the nodes found (Mrs. White and Mrs.Peacock in this instance)
    while(xpathIter.MoveNext( ))
    {
        Console.WriteLine(xpathIter.Current.Value);
    }

    // close the reader.
    reader.Close( );
}

This outputs the following:

Mrs. White
Mrs. Peacock

Discussion

XPath is a very versatile language for performing queries on XML-based data. In order to accomplish our goal, we first created an XML fragment that looks like this:

<?xml version='1.0'?>
<Clue>
    <Participant type='Perpetrator'>Professor Plum</Participant>
    <Participant type='Witness'>Colonel Mustard</Participant>
    <Participant type='Witness'>Mrs. White</Participant>
    <Participant type='Witness'>Mrs. Peacock</Participant>
    <Participant type='Witness'>Mr. Green</Participant>
</Clue>;

We then load this fragment into an XmlTextReader, as shown in Recipe 17.1, then construct an XPathDocument to allow us to create an XPathNavigator, which lets us use XPath syntax to query the XML document shown in the preceding listing. The XmlTextReader reads over the document, checking for well-formedness; the XPathDocument instance wraps the XmlTextReader so we can use XPath to locate nodes (as well as perform XSLT transforms directly), and the XPathNavigator gets the set of nodes selected by the XPath expression.

XmlTextReader reader = new XmlTextReader(xmlFragment, 
                        XmlNodeType.Element,null); 

// Instantiate an XPathDocument using the XmlTextReader.  
XPathDocument xpathDoc = new XPathDocument(reader, XmlSpace.Preserve);
    
// get the navigator
XPathNavigator xpathNav = xpathDoc.CreateNavigator( );

Now we have to determine the XPath-based query to get all of the married female participants who were witnesses. This is set up in the xpathQuery string like this:

// set up the query looking for the married female participants 
// who were witnesses
string xpathQuery = 
  "/Clue/Participant[attribute::type='Witness'][contains(text( ),'Mrs.')]";

In order to get a bit of comprehension of what is going on here, let me explain the syntax a bit:

• /Clue/Participant says "Get all of the Participants under the root level node Clue."

• Participant[attribute::type='Witness'] says "Select only Participants with an attribute called type with a value of Witness."

• Participant[contains(text( ),'Mrs.')] says "Select only Participants with a value that contains `Mrs.'"

Put them all together and we get all of the married female participants who were witnesses.

Once we have an XPathNavigator, we call the Select method on it, passing the XPath- based query to select the nodes we are looking for that are returned via the XPathNodeIterator. We use the XPathNodeIterator to write out the names of the Participants we found and close the XmlTextReader.

See Also

See the "XPathDocument Class," "XPathNavigator Enumeration," and "XPathNodeIterator Class" topics in the MSDN documentation.