Search XDocument using LINQ without knowing the namespace

Question

Is there a way to search an XDocument without knowing the namespace? I have a process that logs all SOAP requests and encrypts the sensitive data. I want to find any elements based on name. Something like, give me all elements where the name is CreditCard. I don't care what the namespace is.

My problem seems to be with LINQ and requiring a xml namespace.

I have other processes that retrieve values from XML, but I know the namespace for these other process.

XDocument xDocument = XDocument.Load(@"C:\temp\Packet.xml");
XNamespace xNamespace = "http://CompanyName.AppName.Service.Contracts";

var elements = xDocument.Root
                        .DescendantsAndSelf()
                        .Elements()
                        .Where(d => d.Name == xNamespace + "CreditCardNumber");

I really want to have the ability to search xml without knowing about namespaces, something like this:

XDocument xDocument = XDocument.Load(@"C:\temp\Packet.xml");
var elements = xDocument.Root
                        .DescendantsAndSelf()
                        .Elements()
                        .Where(d => d.Name == "CreditCardNumber")

This will not work because I don't know the namespace beforehand at compile time.

How can this be done?

<s:Envelope xmlns:s="http://schemas.xmlsoap.org/soap/envelope/">
<s:Body xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema">
    <Request xmlns="http://CompanyName.AppName.Service.ContractA">
        <Person>
            <CreditCardNumber>83838</CreditCardNumber>
            <FirstName>Tom</FirstName>
            <LastName>Jackson</LastName>
        </Person>
        <Person>
            <CreditCardNumber>789875</CreditCardNumber>
            <FirstName>Chris</FirstName>
            <LastName>Smith</LastName>
        </Person>
        ...

<s:Envelope xmlns:s="http://schemas.xmlsoap.org/soap/envelope/">
<s:Body xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema">
    <Request xmlns="http://CompanyName.AppName.Service.ContractsB">
        <Transaction>
            <CreditCardNumber>83838</CreditCardNumber>
            <TransactionID>64588</FirstName>
        </Transaction>      
        ...

Answer 1

As Adam precises in the comment, XName are convertible to a string, but that string requires the namespace when there is one. That's why the comparison of .Name to a string fails, or why you can't pass "Person" as a parameter to the XLinq Method to filter on their name.
XName consists of a prefix (the Namespace) and a LocalName. The local name is what you want to query on if you are ignoring namespaces.
Thank you Adam :)

You can't put the Name of the node as a parameter of the .Descendants() method, but you can query that way :

var doc= XElement.Parse(
@"<s:Envelope xmlns:s=""http://schemas.xmlsoap.org/soap/envelope/"">
<s:Body xmlns:xsi=""http://www.w3.org/2001/XMLSchema-instance"" xmlns:xsd=""http://www.w3.org/2001/XMLSchema"">
  <Request xmlns=""http://CompanyName.AppName.Service.ContractA"">
    <Person>
        <CreditCardNumber>83838</CreditCardNumber>
        <FirstName>Tom</FirstName>
        <LastName>Jackson</LastName>
    </Person>
    <Person>
        <CreditCardNumber>789875</CreditCardNumber>
        <FirstName>Chris</FirstName>
        <LastName>Smith</LastName>
    </Person>
   </Request>
   </s:Body>
</s:Envelope>");

EDIT : bad copy/past from my test :)

var persons = from p in doc.Descendants()
              where p.Name.LocalName == "Person"
              select p;

foreach (var p in persons)
{
    Console.WriteLine(p);
}

That works for me...

Answer 2

You could take the namespace from the root-element:

XDocument xDocument = XDocument.Load(@"C:\temp\Packet.xml");
var ns = xDocument.Root.Name.Namespace;

Now you can get all desired elements easily using the plus-operator:

root.Elements(ns + "CreditCardNumber")

Answer 3

I think I found what I was looking for. You can see in the following code I do the evaluation Element.Name.LocalName == "CreditCardNumber" . This seemed to work in my tests. I'm not sure if it's a best practice, but I'm going to use it.

XDocument xDocument = XDocument.Load(@"C:\temp\Packet.xml");
var elements = xDocument.Root.DescendantsAndSelf().Elements().Where(d => d.Name.LocalName == "CreditCardNumber");

Now I have elements where I can encrypt the values.

If anyone has a better solution, please provide it. Thanks.

Answer 4

There's a couple answers with extension methods that have been deleted. Not sure why. Here's my version that works for my needs.

public static class XElementExtensions
{
    public static XElement ElementByLocalName(this XElement element, string localName)
    {
        return element.Descendants().FirstOrDefault(e => e.Name.LocalName == localName && !e.IsEmpty);
    }
}

The IsEmpty is to filter out nodes with x:nil="true"

There may be additional subtleties - so use with caution.

Answer 5

If your XML documents always defines the namespace in the same node ( Request node in the two examples given), you can determine it by making a query and seeing what namespace the result has:

XDocument xDoc = XDocument.Load("filename.xml");
//Initial query to get namespace:
var reqNodes = from el in xDoc.Root.Descendants()
               where el.Name.LocalName == "Request"
               select el;
foreach(var reqNode in reqNodes)
{
    XNamespace xns = reqNode.Name.Namespace;
    //Queries making use of namespace:
    var person = from el in reqNode.Elements(xns + "Person")
                 select el;
}

Answer 6

I a suffering from a major case of "I know that is the solution, but I am disappointed that that is the solution"... I recently wrote a query like the one below (which I will shortly replace, but it has educational value):

var result = xdoc.Descendants("{urn:schemas-microsoft-com:rowset}data")
.FirstOrDefault()?
.Descendants("{#RowsetSchema}row");

If I remove the namespaces from the XML, I can write the same query like this:

var result = xdoc.Descendants("data")
.FirstOrDefault()?
.Descendants("row");

I plan to write my own extension methods that should allow me to leave the namespaces alone and search for nodes like this:

var result = xdoc.Descendants("rs:data") 
.FirstOrDefault()?
.Descendants("z:row"); 
//'rs:' {refers to urn:schemas-microsoft-com:rowset}
//'z:' {refers to xmlns:z=#RowsetSchema}

My comments just below the code point to how I would like to hide the ugliness of the solution in an Extension Methods library. Again, I'm aware of the solutions posted earlier - but I wish the API itself handled this more fluently . (See what I did there?)

Answer 7

Just use the Descendents method:

XDocument doc = XDocument.Load(filename);
String[] creditCards = (from creditCardNode in doc.Root.Descendents("CreditCardNumber")
                        select creditCardNode.Value).ToArray<string>();