XML：如何删除没有属性和子元素的所有节点

Question

I have a xml document like this :

<Node1 attrib1="abc">
    <node1_1>
         <node1_1_1 attrib2 = "xyz" />
    </ node1_1>
</Node1>

<Node2 />    

Here <node2 /> is the node i want to remove since it has not children/elements nor any attributes.

Answer 1

Using an XPath expression it is possible to find all nodes that have no attributes or children. These can then be removed from the xml. As Sani points out, you might have to do this recursively because node_1_1 becomes empty if you remove its inner node.

var xmlDocument = new XmlDocument();
xmlDocument.LoadXml(
@"<Node1 attrib1=""abc"">
        <node1_1>
             <node1_1_1 />
        </node1_1>
    </Node1>
    ");

// select all nodes without attributes and without children
var nodes = xmlDocument.SelectNodes("//*[count(@*) = 0 and count(child::*) = 0]");

Console.WriteLine("Found {0} empty nodes", nodes.Count);

// now remove matched nodes from their parent
foreach(XmlNode node in nodes)
    node.ParentNode.RemoveChild(node);

Console.WriteLine(xmlDocument.OuterXml);
Console.ReadLine();

Answer 2

Smething like this should do it:

XmlNodeList nodes = xmlDocument.GetElementsByTagName("Node1");

foreach(XmlNode node in nodes)
{
    if(node.ChildNodes.Count == 0)
         node.RemoveAll;
    else
    {
        foreach (XmlNode n in node)
        {
            if(n.InnerText==String.Empty && n.Attributes.Count == 0)
            {
                n.RemoveAll;

            }
        }
    }
}

Answer 3

This stylesheet uses an identity transform with an empty template matching elements without nodes or attributes, which will prevent them from being copied to the output:

<?xml version="1.0" encoding="UTF-8"?>
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">

    <!--Identity transform copies all items by default -->
    <xsl:template match="@* | node()">
        <xsl:copy>
            <xsl:apply-templates select="@*|node()"/>
        </xsl:copy>
    </xsl:template>

    <!--Empty template to match on elements without attributes or child nodes to prevent it from being copied to output -->
    <xsl:template match="*[not(child::node() | @*)]"/>

</xsl:stylesheet>

Answer 4

To do this for all empty child nodes, use a for loop (not foreach) and in reverse order. I resolved it as:

var xmlDocument = new XmlDocument();
xmlDocument.LoadXml(@"<node1 attrib1=""abc"">
                         <node1_1>
                            <node1_1_1 />
                         </node1_1>
                         <node1_2 />
                         <node1_3 />
                      </node1>
                      <node2 />
");
RemoveEmptyNodes(xmlDocument );

private static bool RemoveEmptyNodes(XmlNode node)
{
    if (node.HasChildNodes)
    {
        for(int I = node.ChildNodes.Count-1;I >= 0;I--)
            if (RemoveEmptyNodes(node.ChildNodes[I]))
                node.RemoveChild(node.ChildNodes[I]);
    }
    return 
        (node.Attributes == null || 
            node.Attributes.Count == 0) && 
        node.InnerText.Trim() == string.Empty;
}

The recursive calls (similarly to other solutions) eliminate the duplicated document processing of the xPath approach. More importantly the code is more readable and more readily editable. Win-Win.

So, this solution will remove <node2> , but also correctly removes <node1_2> and <node1_3> .

Update: Found a notable performance increase by using the following Linq implementation.

string myXml = @"<node1 attrib1=""abc"">
                         <node1_1>
                            <node1_1_1 />
                         </node1_1>
                         <node1_2 />
                         <node1_3 />
                      </node1>
                      <node2 />
");
XElement xElem = XElement.Parse(myXml);
RemoveEmptyNodes2(xElem);

private static void RemoveEmptyNodes2(XElement elem)
{
    int cntElems = elem.Descendants().Count();
    int cntPrev;
    do
    {
        cntPrev = cntElems;
        elem.Descendants()
            .Where(e => 
                string.IsNullOrEmpty(e.Value.Trim()) && 
                !e.HasAttributes).Remove();
        cntElems = elem.Descendants().Count();
    } while (cntPrev != cntElems);
}

The loop handles cases where a parent needs to be removed because its only child was removed. Using the XContainer or derivatives tends to have similar performance increases due to the IEnumerable implementations behind the scenes. It's my new favorite thing.

On an arbitrary 68MB xml file RemoveEmptyNodes tends to take about 90sec, while RemoveEmptyNodes2 tends to take about 1sec.