Can MatchCollection in regex find all patterns in C#?

Question

I wrote a simple regex pattern with corresponding sample

   var regex = @"_if_.*_else_.*_endif_";

   // 4 nested regex pattern
   var sample = @"_if_111_else_222_if__if_333_else_444_endif__else_555_if_666_else_777_endif__endif__endif_";

   var matches = Regex.Matches(sample, regex); // count : 1 ?!?!?

Result of matched variables returns only 1 record while I expected it to return 4 records.

_if_666_else_777_endif_
_if_333_else_444_endif_
_if__if_333_else_444_endif__else_555_if_666_else_777_endif__endif__endif_
_if_111_else_222_if__if_333_else_444_endif__else_555_if_666_else_777_endif__endif__endif_

How can I get all patterns that exists in string by regex? Is there a better way?

In which form do you need the results? Do you just want every nested substring that matches the regular expression, or do you want to build some kind of tree? — zneak, Jul 22 '16 at 06:05
@zneak , Exactly this is a nested if so if find correctly should be return 4 matches so I dont know how can I modify it for gain result — user6609534, Jul 22 '16 at 06:07
I mean, what result do you need? Just the number of sub-expressions? Every sub-expression string? Because just every string that matches doesn't tell you where it was to begin with, and that seriously limits your ability to do things. — zneak, Jul 22 '16 at 06:09
I don't think that you can write that with a regular expression. For instance, in `_if__if_111_else_222_endif_else_333_endif_`, I don't know how you would find which `else` goes with the top-level `if` with just a regular expression. — zneak, Jul 22 '16 at 06:41

score 1 · Accepted Answer · answered Jul 22 '16 at 06:48

I suggest a 2 step approach combined into a Regex + Linq.

Get all the balanced substrings from _if_ till _endif_
Only keep those that have _else_ inside.

See IDEONE demo

var s = @"_if_111_else_222_if__if_333_else_444_endif__else_555_if_666_else_777_endif__endif__endif_";
var pat = @"(?x)(?=        # Start of the overlapping match capturing lookahead
         (_if_                 # Leading delimiter
          (?>                  # Start of atomic group (no backtracking into it)
           (?!_(?:end)?if_).   # Any symbol not starting the delimiter sequence 
           |(?<o>_if_)         # A leading delimiter added to stack o
           |(?<-o>_endif_)     # Trailing delimiter added to stack o
          )*                   # Repeat the atomic group 0+ times
          (?(o)(?!))           # If the o stack is not empty, fail the match
         _endif_               # Trailing delimiter
         )
        )";
var res = Regex.Matches(s, pat)
        .Cast<Match>()
        .Select(p => p.Groups[1].Value)
        .Where(n => n.Contains("_else_"))
        .ToList();
foreach (var v in res)
    Console.WriteLine(v);

Can MatchCollection in regex find all patterns in C#?

1 Answers1