Parsing string using regex where delimiter varies

Question

I have the following strings I need to parse:

<!-- MOCK -->
<!--MOCK-->
<!-- MOCK:SENT;FAIL -->
<!--MOCK:SENT;SUCCESS-->
<!--MOCK:SENT; SUCCESS-->
<!--MOCK:SENT ; SUCCESS-->
<!--MOCK:SENT -->
<!-- MOCK:SENT-->
<!-- MOCK : SENT -->
<!-- MOCK: SENT -->
<!-- MOCK :SENT -->

Because these may be entered by a person I want to provide, within reason, support for various spacing.

I've gotten as far as the following regular expression.

<!-- *(?<mode>.*?(?=[ \-:])).*: *(?<responses>.*?(?= *-->))

This works for all of the above test strings except the first two. In those first two cases the "mode" group comes back empty, which I don't understand because I thought the ?<mode>.*?(?=[ \-:]) would collect all characters until it came upon a space, dash, or colon. If I change .*: * in the middle to .*:? * then I get all the "mode" values but I lose all of my "responses" values.

My test code is in .NET Fiddle (https://dotnetfiddle.net/HzIXch), but here it is for completeness.

var regex = new Regex(@"<!-- *(?<mode>.*?(?=[ \-:])).*: *(?<responses>.*?(?= *-->))");
var values = new[] {
    "<!-- MOCK -->", 
    "<!--MOCK-->", 
    "<!-- MOCK:SENT;FAIL -->", 
    "<!--MOCK:SENT;SUCCESS-->", 
    "<!--MOCK:SENT; SUCCESS-->", 
    "<!--MOCK:SENT ; SUCCESS-->", 
    "<!--MOCK:SENT -->", 
    "<!-- MOCK:SENT-->", 
    "<!-- MOCK : SENT -->", 
    "<!-- MOCK: SENT -->", 
    "<!-- MOCK :SENT -->", 
};
foreach (var value in values)
{
    var match = regex.Match(value);
    Console.WriteLine($"Mode: [{match.Groups["mode"].Value}], Responses: [{match.Groups["responses"].Value}]");
}

Make the spaces : '\s*' . The asterisks is zero or more occurrences. — jdweng
– jdweng, Commented Nov 7, 2017 at 1:24
Do your "mode" and "responses" groups always consist of capital letters, or can they contain other characters? — CAustin
– CAustin, Commented Nov 7, 2017 at 1:55

Craig W. · Accepted Answer · 2017-11-07 15:36:04Z

1

I don't think there's any need to use lookahead here. This is how I would do it:

(?<mode>[A-Z]+)([ \-:]+(?<responses>[A-Z; ]+\b))?

https://dotnetfiddle.net/6xpkt1

Regarding the other answer that was posted, and the comment on your question, I completely disagree. \s is not the same as a literal space. It includes all whitespace characters, which can get you into trouble if you're only expecting spaces.

edited Nov 7, 2017 at 15:36

Craig W.

18.2k6 gold badges52 silver badges83 bronze badges

answered Nov 7, 2017 at 1:57

CAustin

4,62315 silver badges27 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Craig W. Over a year ago

When I first checked the fiddle I thought it wasn't working correctly. Then I realized that the regex in the fiddle and this one didn't quite match. When I plugged in the one above it worked correctly. I updated your post to the new fiddle link for future visitors. Looks like I was overthinking it, this is much cleaner than the direction I was going with it. The only change I made was to change [A-Z] to [A-Za-z] in case the user enters lower case.

King King Over a year ago

it's true that the \s can match a space or a tab character. But in almost cases we usually would like to match the tab char as well, so \s is really helpful and exactly what we need here. Even in this case the OP should use \s, I don't think he would accept any tab character as part of a group (mode or response). You seem to get what the OP wants fine but not dare to be sure that he would like to use something like \s instead of literal space in this case ( =)) funny enough).

Collectives™ on Stack Overflow

Parsing string using regex where delimiter varies

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related