This depends on what you want to do, however, if you are primarily interested in the text between two bookmarks, then this is one of those cases where XmlDocument / XPath semantics are easier to use than LINQ to XML or the strongly-typed object model of the Open XML SDK V2. The semantics of the 'following::*' axis of XPath is what you want. The following example uses XmlDocument and XPath to print the names of the nodes between the start and end of a bookmark.
using System;
using System.Collections.Generic;
using System.IO;
using System.Linq;
using System.Text;
using System.Xml;
using DocumentFormat.OpenXml.Packaging;
using DocumentFormat.OpenXml.Wordprocessing;
class Program
{
public static XmlDocument GetXmlDocument(OpenXmlPart part)
{
XmlDocument xmlDoc = new XmlDocument();
using (Stream partStream = part.GetStream())
using (XmlReader partXmlReader = XmlReader.Create(partStream))
xmlDoc.Load(partXmlReader);
return xmlDoc;
}
static void Main(string[] args)
{
using (WordprocessingDocument doc =
WordprocessingDocument.Open("Test.docx", false))
{
XmlDocument xmlDoc = GetXmlDocument(doc.MainDocumentPart);
string wordNamespace =
"http://schemas.openxmlformats.org/wordprocessingml/2006/main";
XmlNamespaceManager nsmgr =
new XmlNamespaceManager(xmlDoc.NameTable);
nsmgr.AddNamespace("w", wordNamespace);
XmlElement bookmarkStart = (XmlElement)xmlDoc.SelectSingleNode("descendant::w:bookmarkStart[@w:id='0']", nsmgr);
XmlNodeList nodesFollowing = bookmarkStart.SelectNodes("following::*", nsmgr);
var nodesBetween = nodesFollowing
.Cast<XmlNode>()
.TakeWhile(n =>
{
if (n.Name != "w:bookmarkEnd")
return true;
if (n.Attributes.Cast<XmlAttribute>().Any(a => a.Name == "w:id" && a.Value == "0"))
return false;
return true;
});
foreach (XmlElement item in nodesBetween)
{
Console.WriteLine(item.Name);
if (item.Name == "w:bookmarkStart" || item.Name == "w:bookmarkEnd")
foreach (XmlAttribute att in item.Attributes)
Console.WriteLine("{0}:{1}", att.Name, att.Value);
}
}
}
}
