Question

How can you do a streaming read on a large XML file that contains a xs:sequence just below root element, without loading the whole file into a XDocument instance in memory?

Answer 1

Going with a SAX-style element parser and the XmlTextReader class created with XmlReader.Create would be a good idea, yes. Here’s a slightly-modified code example from CodeGuru:

void ParseURL(string strUrl) {   try   {     using (var reader = XmlReader.Create(strUrl))     {       while (reader.Read())       {         switch (reader.NodeType)         {           case XmlNodeType.Element:             var attributes = new Hashtable();             var strURI = reader.NamespaceURI;             var strName = reader.Name;             if (reader.HasAttributes)             {               for (int i = 0; i < reader.AttributeCount; i++)               {                 reader.MoveToAttribute(i);                 attributes.Add(reader.Name,reader.Value);               }             }             StartElement(strURI,strName,strName,attributes);             break;             //             //you can handle other cases here             //             //case XmlNodeType.EndElement:             // Todo             //case XmlNodeType.Text:             // Todo             default:             break;           }         }       }       catch (XmlException e)       {         Console.WriteLine('error occured: ' + e.Message);       }     }   } }