Display all possible matches for a regex pattern

Question

I have the following RegEx pattern in order to determine some 3-digit exchanges of phone numbers:

(?:2(?:04|[23]6|[48]9|50)|3(?:06|43|65)|4(?:03|1[68]|3[178]|50)|5(?:06|1[49]|79|8[17])|6(?:0[04]|13|39|47)|7(?:0[59]|78|8[02])|8(?:[06]7|19|73)|90[25])

It looks pretty daunting, but it only yields around 40 or 50 numbers. Is there a way in C# to generate all numbers that match this pattern? Offhand, I know I can loop through the numbers 001 thru 999, and check each number against the pattern, but is there a cleaner, built-in way to just generate a list or array of matches?

ie - {"204","226","236",...}

possible duplicate of [Generate all matches for regex](http://stackoverflow.com/questions/21517428/generate-all-matches-for-regex) — Wiktor Stribiżew, Feb 19 '15 at 22:46
[Here](http://tinyurl.com/m2vj9q9). Just go through all the possible paths. There's nothing built-in. — Lucas Trzesniewski, Feb 19 '15 at 22:49
Thanks @LucasTrzesniewski - that is a cool tool, if only for a visual aid. — WEFX, Feb 20 '15 at 02:35
@WEFX I think the idea is that you can convert a regex to a DFA (unless it's a real fancy regex) and then you can traverse the DFA. — Millie Smith, Feb 20 '15 at 02:40
@Millie exactly. Notice the railroad diagram in my comment above actually *is* a DFA. This specific pattern makes it real easy. — Lucas Trzesniewski, Feb 20 '15 at 08:23

ΩmegaMan · Accepted Answer · 2015-02-20T19:46:40.723

No, there is no off the shelf tool to determine all matches given a regex pattern. Brute force is the only way to test the pattern.

Update

It is unclear why you are using (?: ) which is the "Match but don't capture". It is used to anchor a match, for example take this phone text phone:303-867-5309 where we don't care about the phone: but we want the number.

The pattern used would be

(?:phone\:)(\d{3}-\d{3}-\d{4})

which would match the whole line, but the capture returned would just be the second match of the phone number 303-867-5309.

So the (?: ) as mentioned is used to anchor a match capture at a specific point; with text match text thrown away.

With that said, I have redone your pattern with comments and a test to 2000:

string pattern = @"
^                            # Start at beginning of line so no mid number matches erroneously found
   (
       2(04|[23]6|49|[58]0)  # 2 series only match 204, 226, 236, 249, 250, 280
     |                       # Or it is not 2, then match:
       3(06|43|65)           # 3 series only match 306, 343, 365
    )
$                            # Further Anchor it to the end of the string to keep it to 3 numbers";

// RegexOptions.IgnorePatternWhitespace allows us to put the pattern over multiple lines and comment it. Does not
//     affect regex parsing/processing.

var results = Enumerable.Range(0, 2000) // Test to 2000 so we don't get any non 3 digit matches.
                        .Select(num => num.ToString().PadLeft(3, '0'))
                        .Where (num => Regex.IsMatch(num, pattern, RegexOptions.IgnorePatternWhitespace))
                        .ToArray();

Console.WriteLine ("These results found {0}", string.Join(", ", results));

// These results found 204, 226, 236, 249, 250, 280, 306, 343, 365

@WEFX I updated the answer to give you a different flavor of the pattern/processing. — ΩmegaMan, Feb 20 '15 at 15:16

score 0 · Answer 2 · answered Feb 20 '15 at 14:30

I took the advice of @LucasTrzesniewski and just looped through the possible values. Since I know I’m dealing w/ 3-digit numbers, I just looped through the numbers/strings “000” thru “999” and checked for matches like this:

private static void FindRegExMatches(string pattern)
{
    for (var i = 0; i < 1000; i++)
    {
        var numberString = i.ToString().PadLeft(3, '0');
        if (!Regex.IsMatch(numberString, pattern)) continue;

        Console.WriteLine("Found a match: {0}, numberString);
    }
}

Display all possible matches for a regex pattern

2 Answers2

Update

Linked