Highlight key words in a given search text

Question

Hi, I'm using following method to highlight the keywords in a given text.

private string HighlightSearchKeyWords(string searchKeyWord, string text)
        {
            Regex keywordExp = new Regex(@" ?, ?");
            var pattern = @"\b(" + keywordExp.Replace(Regex.Escape(searchKeyWord), @"|") + @")\b";
            Regex exp = new Regex(pattern, RegexOptions.IgnoreCase | RegexOptions.Singleline);
            return exp.Replace(text, @"<span class=""search-highlight"">$0</span>");

        }

Sample Text: "What is .net Programming? Pl suggest few e-books"

Keyword: ".net"

When i try to search with key word ".net" .net is not getting highlighted in the given sample text.

When i try to search with key word "e-books" e-books is getting highlighted in the given sample text.

What would be the problem. Can anyone pl let me know where exactly do i need modify/

Answer 1

There is no word boundary before ".net" because \b only looks for a change between \w and \W, and both . and (space) fall into the \W category, so there is no boundary between them.

One option is to simply look for "not a word-character" - i.e. not explicitly checking for the boundary, only for the lack of a word character, using a negative lookbehind:

(?<!\w)

You could also check for anything that's not non-whitespace character, like so:

(?<!\S)

This one is a double-negative - it might seem more obvious to do (?<=\s) (or (?<=\W) for previous example), but these will prevent matches at the start of line from matching.

For an example of the difference between these two - the first one would match the .NET in C#.NET whilst the second one would not.

Since you're using .NET regex, you've fortunately got a fairly complete set of regex functionality - but it's worth point out that some other regex implementations don't support negative lookbehind - for those, you would need to use syntax like this:

(?<=\W|^)
(?<=\s|^)

(In all these cases, you want the equivalent lookahead on the other end.)

So, here's how those four variants would look in your pattern:

var pattern = @"(?<!\w)(" + keywordExp.Replace(Regex.Escape(searchKeyWord), @"|") + @")(?!\w)";
var pattern = @"(?<!\S)(" + keywordExp.Replace(Regex.Escape(searchKeyWord), @"|") + @")(?!\S)";
var pattern = @"(?<=\s|^)(" + keywordExp.Replace(Regex.Escape(searchKeyWord), @"|") + @")(?=\s|$)";
var pattern = @"(?<=\W|^)(" + keywordExp.Replace(Regex.Escape(searchKeyWord), @"|") + @")(?=\W|$)";