Question

The title explains it all. It seems simple enough, so I must be overlooking something stupid. Here’s what I’ve got.

    private string getBetween(string strSource, string strStart, string strEnd)
    {
        int start, end;
        if (strSource.Contains(strStart) && strSource.Contains(strEnd))
        {
            start = strSource.IndexOf(strStart, 0) + strStart.Length;
            end = strSource.IndexOf(strEnd, start);
            return strSource.Substring(start, end - start);
        }
        else
        {
            return "";
        }
    }

Thanks, guys.

Answer 1

As couple of peoples said the problem that you code is working on very specific input, it’s all because of this start and end IndexOf magic =) But when you try to update you code to work correct on more inputs you will get into problem that your code become very long with many indexes, comparsions, substrings, conditions and so on. To avoid this i like to recommend you use regular expressions with theirs help you can express what you need on special language.

Here is the sample which solves your problem with regular expressions:

        public static string getBetween(string source, string before, string after)
        {
            var regExp = new Regex(string.Format("{0}(?<needle>[^{0}{1}]+){1}",before,after));
            var matches = regExp.Matches(source).Cast<Match>(). //here we use LINQ to
                OrderBy(m => m.Groups["needle"].Value.Length).  //find shortest string
                Select(m => m.Groups["needle"].Value);          //you can use foreach loop instead
            return matches.FirstOrDefault();
        }

All tricky part is {0}(?<needle>[^{0}{1}]+){1} where 0 – before string and 1 – after string. This expression means that we nned to find string that lies beetween 0 and 1, and also don’t contains 0 and 1.

Hope this helps.