Finding specific attributes in a large XML document

Question

I have a large XML document that is around 100mb. I need to find attributes for two tags in this document. I can do this by using similar code to the following:

XmlDocument xmlDocument = new XmlDocument ( );
xmlDocument.Load ( "C:\\myxml.xml" );

XmlNode node1 = xmlDocument.SelectSingleNode ( "/data/objects[@type='data type 1']" );
if ( null != node1 )
{
   result = node1 [ "Version" ].Value;
}

But doing so loads the entire XML into memory which seems to take around 200mb. Is there anyway I can make this more efficient?

Edit: Lots of nice answers using the XmlTextReader which I have written my code to use now. (It will be more memory efficient, but ugly :).

Answer 1

For performance, SAX is much better than DOM since you actually need only one value. SAX implementation in .NET Framework is XmlTextReader .

Answer 2

You should try to use an XmlReader.

From MSDN :

Like the SAX reader, the XmlReader is a forward-only, read-only cursor. It provides fast, non-cached stream access to the input. It can read a stream or a document. It allows the user to pull data, and skip records of no interest to the application. The big difference lies in the fact that the SAX model is a "push" model, where the parser pushes events to the application, notifying the application every time a new node has been read, while applications using XmlReader can pull nodes from the reader at will.

An example here .

Answer 3

You can use the XmlReader class to do this. A simple but working example that does the same as your code above looks like this:

string result = null;

using (var reader = XmlReader.Create(@"c:\\myxml.xml"))
{
    while (reader.Read())
    {
        if (reader.NodeType == XmlNodeType.Element
            && reader.Depth == 1
            && reader.LocalName == "objects"
            && reader.GetAttribute("type") == "data type 1")
        {
            result = reader.GetAttribute("Version");
            break;
        }
    }
}