Query an XDocument for elements by name at any depth

Question

I have an XDocument object. I want to query for elements with a particular name at any depth using LINQ.

When I use Descendants("element_name"), I only get elements that are direct children of the current level. I'm looking for the equivalent of "//element_name" in XPath...should I just use XPath, or is there a way to do it using LINQ methods?

Polivy

Answer 1

Descendants should work absolutely fine. Here's an example:

using System;
using System.Xml.Linq;

class Test
{
    static void Main()
    {
        string xml = @"
<root>
  <child id='1'/>
  <child id='2'>
    <grandchild id='3' />
    <grandchild id='4' />
  </child>
</root>";
        XDocument doc = XDocument.Parse(xml);

        foreach (XElement element in doc.Descendants("grandchild"))
        {
            Console.WriteLine(element);
        }
    }
}

Results:

<grandchild id="3" />
<grandchild id="4" />

Logogriph

How would you tackle this if an element name was duplicated within an xml document? For example: If the xml contained a collection of <Cars> with sub elements of <Part>, and also a collection of <Planes> with sub elements of <Part>, and you want a list of Parts for Cars only. – Weeds

@pfeds: Then I'd use doc.Descendants("Cars").Descendants("Part") (or possibly .Elements("Part") if they were only direct children. – Logogriph

Six years on and still a fantastic example. In fact, this is still far more helpful than the MSDN explanation :-) – Schild

And it is still an evil example, Dr., since if there are no "Cars", the above code would result in an NPE. Maybe the .? from the new C# will finally make it valid – Blunt

@DrorHarari Nope, no exception is thrown: Try out var foo = new XDocument().Descendants("Bar").Descendants("Baz"); Because Descendants returns an empty IEnumerable<XElement>and not null. – Batson

14 years on and still need this this as example.... – Triglyph

Answer 2

An example indicating the namespace:

String TheDocumentContent =
@"
<TheNamespace:root xmlns:TheNamespace = 'http://www.w3.org/2001/XMLSchema' >
   <TheNamespace:GrandParent>
      <TheNamespace:Parent>
         <TheNamespace:Child theName = 'Fred'  />
         <TheNamespace:Child theName = 'Gabi'  />
         <TheNamespace:Child theName = 'George'/>
         <TheNamespace:Child theName = 'Grace' />
         <TheNamespace:Child theName = 'Sam'   />
      </TheNamespace:Parent>
   </TheNamespace:GrandParent>
</TheNamespace:root>
";

XDocument TheDocument = XDocument.Parse( TheDocumentContent );

//Example 1:
var TheElements1 =
from
    AnyElement
in
    TheDocument.Descendants( "{http://www.w3.org/2001/XMLSchema}Child" )
select
    AnyElement;

ResultsTxt.AppendText( TheElements1.Count().ToString() );

//Example 2:
var TheElements2 =
from
    AnyElement
in
    TheDocument.Descendants( "{http://www.w3.org/2001/XMLSchema}Child" )
where
    AnyElement.Attribute( "theName" ).Value.StartsWith( "G" )
select
    AnyElement;

foreach ( XElement CurrentElement in TheElements2 )
{
    ResultsTxt.AppendText( "\r\n" + CurrentElement.Attribute( "theName" ).Value );
}

Quattlebaum

But, what if my source xml doesn't have a namespace? I suppose I can add one in code (have to look into that), but why is that necessary? In any event, root.Descendants("myTagName") doesn't find elements buried three or four levels deep in my code. – Scarlatti

Thanks! We're using datacontract serialization. This creates a header like <MyClassEntries xmlns:i="w3.org/2001/XMLSchema-instance" xmlns="schemas.datacontract.org/2004/07/DataLayer.MyClass"> and I was stumped why I wasn't getting any descendants. I needed to add the {schemas.datacontract.org/2004/07/DataLayer.MyClass} prefix. – Tormentil

after hours of searching and experimenting, this is the only answer that helped. Man can't thank you enough. Kudos for adding the namespace into the Descendants. – Bogle

Answer 3

You can do it this way:

xml.Descendants().Where(p => p.Name.LocalName == "Name of the node to find")

where xml is a XDocument.

Be aware that the property Name returns an object that has a LocalName and a Namespace. That's why you have to use Name.LocalName if you want to compare by name.

Ize

I'm trying to get all EmbeddedResource node from c# project file, and this is only way that works. XDocument document = XDocument.Load(csprojPath); IEnumerable<XElement> embeddedResourceElements = document.Descendants( "EmbeddedResource"); Is not works and I don't understand why. – Top

Answer 4

Descendants will do exactly what you need, but be sure that you have included a namespace name together with element's name. If you omit it, you will probably get an empty list.

Roundhouse

Answer 5

There are two ways to accomplish this,

1. LINQ to XML
2. XPath

The following are samples of using these approaches,

List<XElement> result = doc.Root.Element("emails").Elements("emailAddress").ToList();

If you use XPath, you need to do some manipulation with the IEnumerable:

IEnumerable<XElement> mails = ((IEnumerable)doc.XPathEvaluate("/emails/emailAddress")).Cast<XElement>();

Note that

var res = doc.XPathEvaluate("/emails/emailAddress");

results either a null pointer, or no results.

Fluxmeter

just to mention that XPathEvaluate is in the System.Xml.XPath namespace. – Ruffian

XPathEvaluate should do the trick, but your query only takes nodes at a particular depth (one). If you wanted to select all elements named "email" regardless of where in a document they occur, you'd use the path "//email". Obviously such paths are more expensive, since the entire tree must be walked whatever the name is, but it can be quite convenient - provided you know what you're doing. – Diversify

Answer 6

I am using XPathSelectElements extension method which works in the same way to XmlDocument.SelectNodes method:

using System;
using System.Xml.Linq;
using System.Xml.XPath; // for XPathSelectElements

namespace testconsoleApp
{
    class Program
    {
        static void Main(string[] args)
        {
            XDocument xdoc = XDocument.Parse(
                @"<root>
                    <child>
                        <name>john</name>
                    </child>
                    <child>
                        <name>fred</name>
                    </child>
                    <child>
                        <name>mark</name>
                    </child>
                 </root>");

            foreach (var childElem in xdoc.XPathSelectElements("//child"))
            {
                string childName = childElem.Element("name").Value;
                Console.WriteLine(childName);
            }
        }
    }
}

Ruffian

Answer 7

This my variant of the solution based on LINQ and the Descendants method of the XDocument class

using System;
using System.Linq;
using System.Xml.Linq;

class Test
{
    static void Main()
    {
        XDocument xml = XDocument.Parse(@"
        <root>
          <child id='1'/>
          <child id='2'>
            <subChild id='3'>
                <extChild id='5' />
                <extChild id='6' />
            </subChild>
            <subChild id='4'>
                <extChild id='7' />
            </subChild>
          </child>
        </root>");

        xml.Descendants().Where(p => p.Name.LocalName == "extChild")
                         .ToList()
                         .ForEach(e => Console.WriteLine(e));

        Console.ReadLine();
    }
}

Results:

For more details on the Desendants method take a look here.

Wolk

Answer 8

Following @Francisco Goldenstein answer, I wrote an extension method

using System.Collections.Generic;
using System.Linq;
using System.Xml.Linq;

namespace Mediatel.Framework
{
    public static class XDocumentHelper
    {
        public static IEnumerable<XElement> DescendantElements(this XDocument xDocument, string nodeName)
        {
            return xDocument.Descendants().Where(p => p.Name.LocalName == nodeName);
        }
    }
}

Silvia

Answer 9

We know the above is true. Jon is never wrong; real life wishes can go a little further.

<ota:OTA_AirAvailRQ
    xmlns:ota="http://www.opentravel.org/OTA/2003/05" EchoToken="740" Target=" Test" TimeStamp="2012-07-19T14:42:55.198Z" Version="1.1">
    <ota:OriginDestinationInformation>
        <ota:DepartureDateTime>2012-07-20T00:00:00Z</ota:DepartureDateTime>
    </ota:OriginDestinationInformation>
</ota:OTA_AirAvailRQ>

For example, usually the problem is, how can we get EchoToken in the above XML document? Or how to blur the element with the name attribute.

1. You can find them by accessing with the namespace and the name like below

    doc.Descendants().Where(p => p.Name.LocalName == "OTA_AirAvailRQ").Attributes("EchoToken").FirstOrDefault().Value
   

2. You can find it by the attribute content value, like this one.

Sticktight

Answer 10

(Code and Instructions is for C# and may need to be slightly altered for other languages)

This example works perfect if you want to read from a Parent Node that has many children, for example look at the following XML;

<?xml version="1.0" encoding="UTF-8"?> 
<emails>
    <emailAddress>[email protected]</emailAddress>
    <emailAddress>[email protected]</emailAddress>
    <emailAddress>rgreen@set_ig.ca</emailAddress> 
</emails>

Now with this code below (keeping in mind that the XML File is stored in resources (See the links at end of snippet for help on resources) You can obtain each email address within the "emails" tag.

XDocument doc = XDocument.Parse(Properties.Resources.EmailAddresses);

var emailAddresses = (from emails in doc.Descendants("emailAddress")
                      select emails.Value);

foreach (var email in emailAddresses)
{
    //Comment out if using WPF or Windows Form project
    Console.WriteLine(email.ToString());

   //Remove comment if using WPF or Windows Form project
   //MessageBox.Show(email.ToString());
}

Results

1. [email protected]
2. [email protected]
3. rgreen@set_ig.ca

Note: For Console Application and WPF or Windows Forms you must add the "using System.Xml.Linq;" Using directive at the top of your project, for Console you will also need to add a reference to this namespace before adding the Using directive. Also for Console there will be no Resource file by default under the "Properties folder" so you have to manually add the Resource file. The MSDN articles below, explain this in detail.

Adding and Editing Resources

How to: Add or Remove Resources

Leah

Don't want to be mean here, but your example does not show grandchildren. emailAddress is a child of emails. I am wondering if there is a way to use Descendants without using namespaces? – Philanthropy