Earlier today, a colleague of mine came to me with a question. He was writing Unit Tests for a serialization utility and needed to compare the generated xml with the hand written xml file which was the expected result.

By eye inspection, the xml seemed to be the same, but just in a different order. The elements were in a different order and the attributes were in a different order. So although order didn’t matter functionally, it eliminated the ability to do:docA.InnerXml.Equals(docB.InnerXml)

After some googling, I found there were a few tools out there to do XML comparison, but they were bloated with too many options and features. I didn’t want to have to learn something new; I just wanted to see if the content of the xml was the same regardless of order.

So, about an hour later, I came up with my own version that had no learning curve and no extra options. It just had one public method: AreEqualNoOrder.

It seemed like this kind of thing could be useful to a lot of other people so I decided to post it. Attached is a console application that runs a few tests on the comparer to get a better understanding of what it does.

using System.Xml;

namespace XMLComparer
{
    /// <summary>
    /// Compares two xml files for equality of content in any order.
    /// </summary>
    public class XmlComparer
    {
        /// <summary>
        /// Compares two xml documents for eqality as defined for this comparer.
        /// </summary>
        /// <param name="docA">The first xml document.</param>
        /// <param name="docB">the second xml document.</param>
        /// <returns>True if the documents are equal, otherwise false.</returns>
        public static bool AreEqualNoOrder(XmlDocument docA, XmlDocument docB)
        {
            return nodesAreEqualNoOrder(docA.FirstChild, docB.FirstChild);
        }

        /// <summary>
        /// Compares two nodes for equality as defined for this comparer. 
        /// </summary>
        /// <param name="nodeA">A node from the first document.</param>
        /// <param name="nodeB">a node from the second document.</param>
        /// <returns>True if the nodes are equal, otherwise false.</returns>
        private static bool nodesAreEqualNoOrder(XmlNode nodeA, XmlNode nodeB)
        {
            ///////////////
            // Compare Text
            ///////////////
            var textA = nodeA.Value;
            var textB = nodeB.Value;

            if (textA == null || textB == null)
            {
                // if either is null, then they should both be null.
                if (!(textA == null && textB == null))
                {
                    return false;
                }
            }
            else
            {
                // if they are not null, the text should be the same
                if (!textA.Trim().Equals(textB.Trim()))
                {
                    return false;
                }
            }

            /////////////////////
            // Compare Attributes
            /////////////////////
            var attributesA = nodeA.Attributes;
            var attributesB = nodeB.Attributes;

            if (attributesA == null || attributesB == null)
            {
                // if either is null, then they should both be null.
                if (!(attributesA == null && attributesB == null))
                {
                    return false;
                }
            }
            else
            {
                // if there are attributes, there should be the same number on A as on B.
                if (attributesA.Count != attributesB.Count)
                {
                    return false;
                }

                // check each attribute and value
                for (int i = 0; i < attributesA.Count; i++)
                {
                    var name = attributesA[i].Name;

                    // if nodeA has an attribute named x, then so should nodeB.
                    if (attributesB[name] == null)
                    {
                        return false;
                    }

                    // if nodeA and nodeB both have attributes named x, they should have the same value.
                    if (attributesB[name].Value != attributesA[name].Value)
                    {
                        return false;
                    }
                }
            }

            //////////////////////
            // Compare Child Nodes
            //////////////////////
            var childsA = nodeA.ChildNodes;
            var childsB = nodeB.ChildNodes;

            // the number of children of nodeA should be the same as the number of childeren of nodeB.
            if (childsA.Count != childsB.Count)
            {
                return false;
            }

            // every child of nodeA should have a matching child of nodeB, but not necessarily in the same order.
            while (childsA.Count > 0)
            {
                // look for a match in nodeB for the first child of nodeA
                var matchFound = false;
                for (int i = 0; i < childsB.Count; i++)
                {
                    if (nodesAreEqualNoOrder(childsA[0], childsB[i]))
                    {
                        // if a match is found in nodeB, remove the child in nodeA and the match in nodeB, then move on.
                        // this is important because if A had child nodes x and x and B had child nodes x and y,
                        // then both A nodes would match, but B would have a y that wasn't found.
                        matchFound = true;
                        childsA[0].ParentNode.RemoveChild(childsA[0]);
                        childsB[i].ParentNode.RemoveChild(childsB[i]);
                        break;
                    }
                }
                if (!matchFound)
                {
                    // if no match is found, the nodes ain't the same.
                    return false;
                }
            }

            // if no test failed, the nodes are equal.
            return true;
        }

    }
}

Recommended Answers

All 4 Replies

Bug on line 69, what if there are no attributes of B? Change it to read

if (attributesB == null || attributesA.Count != attributesB.Count)

Thanks Momerath! Fixed now. The problem was also on line 35.

Neat tool, could come in pretty handy. And the recursive approach is well thought out.

But why not write the xml file the way that you want it rather than making a tool to fix this problem?

The problem with making the xml the way we want it is that if the thing we're testing that generates xml changes the order of the attributes, it doesn't really matter, but the test would break and require work.

True, how many time is it going to happen...? Maybe never, but if you have the comparer, why not use it in case?

Be a part of the DaniWeb community

We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.