Why is For loop on Regex Matches slow?

Question

I have the following code:

        string pattern = @"(?:\S+\s){1,6}\S*" + search + @"\S*(?:\s\S+){1,6}";
        String dbContents = row[2].ToString();
        var matches = Regex.Matches(dbContents, pattern, RegexOptions.IgnoreCase | RegexOptions.Compiled);
        for (int i = 0; i < matches.Count; i++)
        {
            if (i == 3)
                break;

            Contents += String.Format("... {0} ...", matches[i].Value);
        } 

What I'm trying to accomplish is to get one to six words before the search term and 1-6 words after the search term. When executing the code the performance hit on the for loop "matches.Count". With very large strings, its taking upwards of a min to execute. I'm confused on why and what to do to fix the issue.

Answer 1

In order to find the count, that has to find all the matches in order to count them. Given that you're going to stop after three anyway, that seems a little pointless.

Use MatchCollection's lazy evaluation in combination with the Take method from LINQ to only take the first three matches. Usually it's a good idea to use StringBuilder rather than string concatenation in a loop, too:

StringBuilder builder = new StringBuilder(...);
foreach (var match in matches.Cast<Match>().Take(3))
{
    builder.AppendFormat("... {0} ...", matches[i].Value);
}

(The StringBuilder change probably isn't going to make much difference here, but it's a good habit to get into. The Cast method is required because Enumerable.Take only works on the generic IEnumerable<T> type.)