I have an application written in C# and i am using MS XML
DOM!

I have a document with the following structure (only the  
<DicEntry> - Elements are important):

<NewDataSet xmlns:xsi="http://www.w3.org/2001/XMLSchema-
instance">
 .
 <DicEntry TextKey="Test1" KeyDependency="True">
   <default>Test1</default>
   <de>Button Germanmanuell</de>
   <it />
 </DicEntry>
 <DicEntry TextKey="Dialog1" KeyDependency="True">
   <default>Dialog1</default>
   <de />
   <it />
 </DicEntry>
 .
</NewDataSet>

I now want to merge such a document with a document of
type XmlDataDocument with the same structure in memory!

There is the main code:

XmlElement root = this.DictionaryDocument.DocumentElement;
//bei Formbezogene "TextKey" = "default"

ls = @"/*/DicEntry[@TextKey=" +
TTUtility.GetXPathConcatParamList(this.TextKey) + "]
[@KeyDependency=" + TTUtility.GetXPathConcatParamList
(this.CurrentMergeNode.Attributes.GetNamedItem
("KeyDependency").Value.ToString()) + "]/*[name()=" +
TTUtility.GetXPathConcatParamList(mNode.Name) + "]";

lNode = root.SelectSingleNode(ls);

The method GetXPathConcatParamList builds a parameterlist
for the Xpath method concat.

/// <summary>
   /// this method builds a parameterlist for the Xpath
method concat,
   /// so that the method SelectSingleNode can also work
(search) on
   /// elements, attributes with double quotation marks
in their names.
   /// </summary>
   /// <param name="aValue">a string</param>
   /// <returns>concat condition for SelectSingleNode
method</returns>
   public static string GetXPathConcatParamList(string
aValue)
   {
     char[] lSplit = new char[1] {'"'};
     string[] lSplitValues={""};
     int lSplitValuesLen;
     string lReturn="";
     string lValue="";
     string lHelp="";

     //replace &quot; in aValue with "
     lValue = aValue.Replace("&quot;", "\"");

     //split with "
     lSplitValues = lValue.Split(lSplit);
     lSplitValuesLen = lSplitValues.Length;
     //if there is no double quotation mark then return
     //and there is no build for concat
     if(lSplitValuesLen == 1)
     {
       lReturn = "\"" + aValue + "\"";
     }
     else //build parameterlist for method concat
     {
       for (int i = 0; i <= lSplitValuesLen - 1; i++)
       {
         if(i==lSplitValuesLen - 1)
           lHelp = lHelp + "\"" + lSplitValues[i] + "\"";
         else
           lHelp = lHelp + "\"" + lSplitValues[i] + "\""
+ "," + "'\"'" + ",";
       }
       lReturn = "concat(" + lHelp + ")";
     }
     return lReturn;
   }

Performance results:

Merging   10 <DicEntry> - Elements into       100      in
memory =>     ca.     1 sec.
Merging   10 <DicEntry> - Elements into       500      in
memory =>     ca.     6 sec.
Merging   10 <DicEntry> - Elements into     1000      in
memory =>     ca.     12 sec.
Merging  100 <DicEntry> - Elements into       100      in
memory =>     ca.     21 sec.
Merging  100 <DicEntry> - Elements into     1000      in
memory =>     ca.   190 sec.
Merging 1000 <DicEntry> - Elements into     1000      in
memory =>     ca.   over 3600 sec.

The method "SelectSingleNode" takes time of 96 % of the
applications running time!!!!!!!!!!!!!!!!!!!!!

Please can you help on this problem?

thanks,
john