Question

I need a regular expression to get a word from a text input based on a index. Here the word should be in flowe brackets.

I will try to explain by giving an example.
If my input is hh{jhh}jkhjh{jkjhh{kljk}j}adajskjj}. For any character index, if it lies in between a pair of brackets I have to fetch that word including brackets.
In the above input if index is 2 or 4 output should be {jhh}.

Answer 1

Try this. There’s a number of problems, of course. First, you didn’t say how you want to handle nested brackets. If you need to handle nested brackets, you’re out of luck with a simple regular expression, because you need a context-free grammar for that. Also note that the following implementation is naive in that it has linear time complexity on a word lookup (time to lookup a word is directly proportional to the number of matches found). That would only be a problem for large number of matches, however. Better implementation could use, for instance, a sorted array of matches and binary search for lookup.

using System.Text.RegularExpressions;

namespace WindowsFormsApplication1
{
    public class ParsedInput
    {
        public static readonly Regex RegularExpression = new Regex("{[a-z]*}");

        private MatchCollection matches;

        public ParsedInput(string input)
        {
            matches = RegularExpression.Matches(input);
        }

        public bool TryGetWord(int index, out string word)
        {
            foreach (Match match in matches)
            {
                if (index >= match.Index && index < (match.Index + match.Length))
                {
                    word = match.Captures[0].Value;
                    return true;
                }
            }

            word = "";
            return false;
        }
    }
}

To use, do:

var parsed = new ParsedInput(input);
string word = "";

if (parsed.TryGetWord(index, out word))
{
    // A word exists at the given index, do something.
}
else
{
    // Handle missing word at the given index (optional).
}