C# Removing entire elements from XMLDocument based on list of unwanted attribute values

Question

I have an XML file, e.g.

<Bars1>
    <Bar name='0'>245</Bar>
    <Bar name='1'>180</Bar>
    <Bar name='2'>120</Bar>
    <Bar name='3'>60</Bar>
    <Bar name='4'>0</Bar>
</Bars1>
<Bars2>
    <Bar name='0'>25</Bar>
    <Bar name='1'>10</Bar>
    <Bar name='2'>10</Bar>
    <Bar name='3'>6</Bar>
    <Bar name='4'>0</Bar>
</Bars2>
<Gubbins3>
    <Bar name='0'>45</Bar>
    <Bar name='1'>18</Bar>
    <Bar name='2'>12</Bar>
    <Bar name='3'>4</Bar>
    <Bar name='4'>0</Bar>
</Gubbins3>

and a List<int> notNeededBarNames, containing e.g. { 1, 3 }

I have loaded the XML file into an XmlDocument xmlDoc, and want to remove ANY "Bar" element where the attribute "name" is one of the integers from my list, regardless of where it might exist in the XML . My example is small, but in reality the document and the list could be quite large.

Is there a nice elegant approach to doing this? I can "brute force" it, but I can't help feeling there might be a better way.

Hope you can help!

Answer 1

Linq2Xml can make the life easier.

var notNeededBarNames = new List<int>() { 1, 3 };
var xDoc = XDocument.Load(filename);

xDoc.Descendants("Bar")
    .Where(bar => notNeededBarNames.Contains( (int)bar.Attribute("name")) )
    .Remove();

var newXml = xDoc.ToString();

Answer 2

Something like this will work

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

class Program
{
    static void Main()
    {
        var notNeededBarNames = new[] {1, 3};
        var elems = new[] {"Bars1", "Bars2", "Gubbins3"};
        var newXDoc = new XElement("root");
        var xDoc = XDocument.Load("PathToYourXmlFile");
        foreach (var elem in elems)
        {
            newXDoc.Add(new XElement(elem));
            var curElem = newXDoc.Elements(elem).Single();
            string xpath = String.Format("root/{0}/Bar", elem);
            var childElems = xDoc.XPathSelectElements(xpath);
            foreach (var childElem in childElems)
            {
                bool add = true;
                var nameAtt = childElem.Attribute("name");
                if (nameAtt != null)
                {
                    int val = Convert.ToInt32(nameAtt.Value);
                    if (notNeededBarNames.Any(x => x == val))
                    {
                        add = false;
                    }
                }
                if (add)
                {
                    curElem.Add(childElem);   
                }
            }
        }
        var newXml = newXDoc.ToString();
    }
}

Answer 3

For an xpath-based solution, you can select the nodes to remove as follows:

private static void TestXmlBars()
    {
            string s = @"<root>
                                        <Bars1>...</Bars1>
                                        <Bars2>...</Bars2>
                                        <Gubbins3>...</Gubbins3>
                                    </root>";

            XmlDocument doc = new XmlDocument();
            doc.LoadXml(s);

            List<int> exclude = new List<int>{1,2};
            // create comma-delimited list
            string list = "," + String.Join(",", exclude) + ","; 
            string xpath = String.Format("/root/*/Bar[contains('{0}', concat(',', @name, ','))]", list);
            XmlNodeList nodesToExclude = doc.SelectNodes(xpath);
            foreach (XmlNode node in nodesToExclude)
            {
                    node.ParentNode.RemoveChild(node);
            }
            Console.WriteLine(doc.OuterXml);
    }

(Note that xpath 1 is constrained as it has no 'in' functionality, hence the 'contains' workaround - see XPath 1.0 to find if an element's value is in a list of values)