Kirk Evans Blog

.NET From a Markup Perspective

Reading Fragments with XmlTextReader

Reading Fragments with XmlTextReader

  • Comments 3

Daniel points out a way to read fragments with XmlTextReader, and mentions that I posed the problem.

I was working with some rather nasty “XML” that was not well-formed, and had to find a way to successfully parse it.  The SgmlReader proved perfect for this, save for 1 outstanding problem:  the XML did not contain a root node.  SgmlReader will read the first node in a document until it encounters the closing element for the first node encountered.  When this happens, it sets its internal state to EOF (End of File).

Using Daniel's approach, I could “trick” the SgmlReader to thinking that there was a root node for the document, and then parse the entire document as though it were well-formed.

FileStream fs = new XmlFragmentStream(“someBIGfile.xml“);
SgmlReader reader = new Sgml.SgmlReader();  
reader.InputStream = new StreamReader(fs,System.Text.Encoding.UTF8);
reader.WhitespaceHandling = System.Xml.WhitespaceHandling.None;
object docNode = reader.NameTable.Add("DOCUMENT");

while(reader.Read() )
 //Use the NameTable for reference comparison                    
 if(reader.NodeType == XmlNodeType.Element)
  if(docNode == reader.Name)

  • Good to know the complete scenario. Now it makes even more sense the approach at the stream level as compared to what Oleg suggested about using the special XmlTextReader constructor taking the XmlParsingContext stuff...
  • Take Outs for 26 April 2004
Page 1 of 1 (3 items)
Leave a Comment
  • Please add 6 and 4 and type the answer here:
  • Post
Translate This Page