Extract all nodes with specific Name, but without nested elements

How could I extract all the nodes from the current node, but without nested elements?

For example this could be my document:

<Document>
    <Members>
        <Member Name="A">
            <Member Name="A"/>
            <Member Name="B"/>
        </Member>
        <Member Name="B"/>
        <Member Name="C">
            <Member Name="A">
                <Member Name="A"/>
            </Member>          
        </Member>
    </Members>
</Document> 

I could use the following:

var nodes = xml.SelectNodes(".//Member");

but it results with selecting all the nodes, even nested.

I would like to use it recurrently to extract nested nodes, so I cannot write:

var nodes = xml.SelectNodes("Document/Members/Member");

So basically, what I would like to receive is:

• when I call it on members, it returns members A,B,C

• when I call it on member A, it returns nested members A,B

• when I call it on member C, it returns nested member A

XPath 3.1 has an innermost function so I think you want e.g. innermost(expression//Member). XPath 3.1 is supported for .NET/C# using Saxon-HE .NET.

With XPath 1.0 as supported by Microsoft I think e.g. expression//MEMBER[not(node())] will give the leaf MEMBER elements.

You need to use an XQuery predicate filter to get only the Member nodes with the right @Name attribute, then go down to all its Member nodes.

var someName = "A";

var nodes = xml.SelectNodes($@"//Member[@Name=""{someName}""]/Member");

For the first requirement, you would just do

var nodes = xml.SelectNodes($@"//Members/Member");

ideally you should use the newer XDocument and XNode classes, as they are faster and easier to use.

The double slash indicates the descendant path, whereas it sounds like you want the child path. This is ./Member or more simply Member.

Bear in mind though that the child elements your get back still have their descendant elements within them (i.e. you've only selected the children, but the children contain their descendants).

EDIT

From your comment, I'm not sure I fully understand the requirement, but if want to use a single expression (i.e. one call to SelectNodes) but don't want the leaf nodes, you could try:

//*[Member]

which is asking for all elements that have a child element called Member.

Maybe there is an easier way to do this but with using Linq I found it to be easy:

var members = XElement.Parse(myXML).Element("Members");
Func<IEnumerable<XElement>, IEnumerable<XElement>> NoNested = (x) => {return x.Select(n => new XElement("Member", new XAttribute("Name", (string)n.Attribute("Name"))));};

var xml = XElement.Parse(myXML);
var selABC = NoNested(xml.XPathSelectElements("./Members/Member"));
var selA = NoNested(xml.XPathSelectElements("./Members/Member[@Name='A']/Member"));
var selC = NoNested(xml.XPathSelectElements("./Members/Member[@Name='C']/Member"));

Here is Dotnet Fiddle Sample