How do I extract the *immediate* text from a C# System.Windows.Form.HtmlElement (i.e. NOT the text in children)

Question

In C#, how do I get the text of an System.Windows.Form.HtmlElement not including the text from its children?

If I have

<div>aaa<div>bbb<div>ccc</div><div>ddd</div></div></div>

then the InnerText property of the whole thing is "aaabbbcccddd" and I just want "aaa".

I figure this should be trivial, but I haven't found anything to produce the "immediate" text of an HtmlElement in C#. More ludicrous ideas are "subtracting" the InnerText of the children from the parent, but that's an insane amount of work for something that I'm sure is trivial.

(All I want is access to the Text Node of the HtmlElement.)

I'd certain appreciate any help (or pointer) that anyone can supply.

Many thanks.

Examples:

<div>aaa<div>bbb<div>ccc</div><div>ddd</div></div></div>  -> Produce "aaa"
<div><div>ccc</div><div>ddd</div></div>                   -> Produce ""
<div>ccc</div>                                            -> Produce "ccc" 

Edit

There are a number of ways to skin this particular cat, none of them elegant. However, given my constraints (not my HTML, quite possibly not valid), I think Aleksey Bykov's solution is closest to what I needed (and indeed, I did implement the same solution he suggested in the last comment.)

I've selected his solution and upvoted all the other ones that I think would work, but weren't optimal for me. I'll check back to upvote any other solutions that seem likely to work.

Many thanks.

Answer 1

Maybe it's simpler than that, if you're willing to use XmlDocument instead of HtmlDocument - you can just use the 'Value' property of the XmlElement.

This code gives the output you want for the 3 cases you mentioned:

class Program
{
    private static string[] htmlTests = {@"<div>aaa<div>bbb<div>ccc</div><div>ddd</div></div></div>",
                                         @"<div><div>ccc</div><div>ddd</div></div>",
                                         @"<div>ccc</div>" };
    static void Main(string[] args)
    {
        var page = new XmlDocument();

        foreach (var test in htmlTests)
        {
            page.LoadXml(test);
            Console.WriteLine(page.DocumentElement.FirstChild.Value);
        }
    }
}

Output:

aaa

ccc

Answer 2

I am not sure what you mean by HtmlElement, but with XmlElement you would do it like this:

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

public static class XmlUtils {

    public static IEnumerable<String> GetImmediateTextValues(XmlNode node) {
        var values = node.ChildNodes.Cast<XmlNode>().Aggregate(
            new List<String>(),
            (xs, x) => { if (x.NodeType == XmlNodeType.Text) { xs.Add(x.Value); } return xs; }
        );
        return values;
    }

    public static String GetImmediateJoinedTextValues(XmlNode node, String delimiter) {
        var values = GetImmediateTextValues(node);
        var text = String.Join(delimiter, values.ToArray());
        return text;
    }
}

EDIT:

Well, if your HtmlElement comes from System.Windows.Forms, then what you need to do is to use its DomElement property trying to cast it to one of the COM interfaces defined in mshtml. So all you need to do is to be able to tell if the element you are looking at is a text node and get its value. First you gotta add a reference to the mshtml COM library. You can do something like this (I cannot verify this code immediately).

public Bool IsTextNode(HtmlElement element) {
  var result = false;
  var nativeNode = element.DomElement as mshtml.IHTMLDOMNode;
  if (nativeNode != null) {
      var nodeType = nativeNode.nodeType;
      result = nodeType == 3; // -- TextNode: http://msdn.microsoft.com/en-us/library/aa704085(v=vs.85).aspx
  }
  return result

}

Answer 3

Well, you could do something like this (assuming your input is in a string called `input'):

string pattern = @">.*?<";
Regex rgx = new Regex(pattern, RegexOptions.IgnoreCase); 

MatchCollection matches = rgx.Matches(input);
var first_match = matches[0].ToString();
string result = first_match.Substring(1, first_match.Length - 2);

I probably wouldn't do it (or just relay on matching the string for the first <div> and </div>) ... here, for extra credit:

int start = pattern.IndexOf(">") + 1;
int end = pattern.IndexOf("<", start);
string result = input.Substring(start, end - start);