Regular expression for treeview

Question

I've try to parse string like this:

"#1#Process#{some process|info}{some name|some info {child info|child info}}{some name|some info}"

in several messages and create string like this:

#1#Process#
-some process|info
-some name|some info
   -child info|child info
-some name|some info

I'm trying to use RegExp and following code:

using System;
using System.Collections;
using System.Text.RegularExpressions;

namespace prRegEXP
{
    class Program
    {
        static String st="";

        public static void Main(string[] args)
        {
            Console.WriteLine("Hello World!");

            // TODO: Implement Functionality Here
            var pattern = @"\{(.*?)\}";
            var query = "#1#Process#{some process|info}{some name|some info {child info|child info}}{some name|some info}";
            FindTree (pattern, query);
            Console.WriteLine(st);
            Console.WriteLine();
            Console.WriteLine("Press any key to continue . . . ");
            Console.ReadKey(true);
        }       


        private static void FindTree (String pattern, String query) {
            var matches = Regex.Matches(query, pattern);            
            foreach (Match m in matches) {
                st += m.Groups[1] + "\n";

                if (Regex.IsMatch(m.Groups[1].ToString(), @"\{(.*?)" )) {                   
                    FindTree (@"\{(.*?)", m.Groups[1].ToString());

                }    
            }
        }
    }
}

It's based on example solution I found and I want to create some message tree which take care about messages inside (like child info|child name). And there can be a lot of them.

I cannot figure out how to match child expressions and send it in recursive parameter. Has any idea or fix?

You could achieve that by using recursive regexp, but it'd require good understanding to work efficiently (and even if supported by language - retrieval of recursive capturing groups might be tricky). Without it - You can construct a 2-depth pattern for Your 2-depth examples, but it's MUCH easier to just write a proper tokenizer with depth-counter... — Vlad, Feb 26 '14 at 11:33

score 2 · Answer 1 · answered Feb 26 '14 at 12:02

Writing a simple regex to support N-depth recursions would be impossible(?) or at least very difficult.

A much easier solution would be to just go trough the string char by char and insert indentation and newlines when a new message is found.

Something along the lines of this should work:

private static String FindTree(String query)
    System.Text.StringBuilder sb = new System.Text.StringBuilder();
    String indent = "";
    foreach (var ch in query)  {
        if (ch == '{') {
            sb.Append("\n");
            sb.Append(indent);
            sb.Append("- ");
            indent += "\t";
        } else if (ch == '}') {
            indent = indent.Substring(1);
        } else {
            sb.Append(ch);
        }
    }
    return sb.ToString();
}

The above code is not tested, nor am I well versed in C# so it might be full of errors, but it should illustrate the basic idea.

@Kaha, RE is not a good instrument to do complete parsing, though You can use it efficiently during parsing for tokenizing, e.g. something like `([{\|}]|[^{\|}]+)` (not sure on C# escapes) can improve approach suggested by *rvalvik* from char-by-char up into string-by-string mode. — Vlad, Feb 26 '14 at 17:01
Thanks to all of you. We already had solution without RE, so I was wondering about using RE. — Kaha, Feb 27 '14 at 03:13

Regular expression for treeview

1 Answers1