.NET & XML

When to Use XPath

You should use XPath when you have an XML node in memory and you wish to navigate directly to a particular child node. This presumes that you have either created or loaded an XmlDocument in memory. You can also load an XML document directly into an XPathDocument from a Stream, URL, TextReader, or XmlReader. This method obviates the need to create an XmlDocument at all, and is more efficient than the DOM, since the XPathDocument is a read-only representation of the XML document.

XPath is a good substitute for XmlReader when you have already read an entire document into memory, and the document is to be processed randomly. If you have an extremely large XML document, or you wish to access it strictly sequentially, however, there can be a performance advantage to writing an XmlReader client that handles parsing events. For example, if you are only interested in a certain node within the document, there is no need to load the entire document into memory; you should write an XmlReader client to handle the specific parsing event that indicates the node in question has been read, and skip the rest.