Parsing string - with regex or something similar?

Question

I'm writing routing class and need help. I need to parse $controller variable and assign parts of that string to another variables. Here is examples of $controller:

$controller = "admin/package/AdminClass::display"
//$path = "admin/package";
//$class = "AdminClass";
//$method = "display";

$controller = "AdminClass::display";
//$path = "";
//$class = "AdminClass";
//$method = "display";

$controller = "display"
//$path = "";
//$class = "";
//$method = "display";

This three situations is all i need. Yes, i can write long procedure to handle this situations, but what i need is simple solution with regex, with function preg_match_all

Any suggestion how to do this?

Please be more spcific in your question regarding other parsing situations / other strings you may need to parse. First to second value could be done via replace/substring/split. — Smamatti
– Smamatti, Commented Dec 22, 2011 at 20:30
@CanSpice, to match this three situations, and then extract parts of string to variables — sasa
– sasa, Commented Dec 22, 2011 at 20:40

Andrew Clark · Accepted Answer · 2011-12-22 20:39:47Z

5

The following regex should accomplish this for you, you can then save the captured groups to $path, $class, and $method.

(?:(.+)/)?(?:(.+)::)?(.+)

Here is a Rubular: http://www.rubular.com/r/1vPIhwPUub

Your php code might look something like this:

$regex = '/(?:(.+)\/)?(?:(.+)::)?(.+)/';
preg_match($regex, $controller, $matches);
$path = $matches[1];
$class = $matches[2];
$method = $matches[3];

edited Dec 22, 2011 at 20:39

answered Dec 22, 2011 at 20:33

Andrew Clark

210k36 gold badges285 silver badges310 bronze badges

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

3 Comments

sasa Over a year ago

Thanks! If you have any link where I can easy learn regex I will be very grateful :)

Andrew Clark Over a year ago

I use regular-expressions.info/reference.html as a cheat-sheet, and that website has some good tutorials. Rubular is also really useful for quickly testing your regex.

fge Over a year ago

This will also match a/b, which is not a legal input!

fge · Accepted Answer · 2011-12-22 21:52:46Z

3

This supposes that paths within the class, and the method name, can only contain letters.

The full regex is the following:

^(?:((?:[a-zA-Z]+/)*)([a-zA-Z]+)::)?([a-zA-Z]+)$

Two non capturing groups: the first one which makes all the path and class optional, the second which avoids the capture of individual path elements.

Explanation:

a path element is one or more letters followed by a /: [a-zA-Z]+/;
there may be zero or more of them: we must apply the * quantifier to the above; but the regex is not an atom, we therefore need a group. As we do not want to capture individual path elements, we use a non capturing group: (?:[a-zA-Z]+/)*;
we want to capture the full path if it is there, we must use a capturing group over this ((?:[a-zA-Z]+/)*);
the method name is one or more letters, and we want to capture it: ([a-zA-Z]+);
if present, it follows the path, and is followed by two semicolons: ((?:[a-zA-Z]+/)*)([a-zA-Z]+)::;
but all this is optional: we must therefore put a group around all this, which again we do not want to capture: (?:((?:[a-zA-Z]+/)*)([a-zA-Z]+)::)?;
finally, it is followed by a method name, which is NOT optional this time, and which we want to capture: (?:((?:[a-zA-Z]+/)*)([a-zA-Z]+)::)?([a-zA-Z]+);
and we want this to match the whole line: we need to anchor it both at the beginning and at the end, which gives the final result: ^(?:((?:[a-zA-Z]+/)*)([a-zA-Z]+)::)?([a-zA-Z]+)$

Phew.

edited Dec 22, 2011 at 21:52

answered Dec 22, 2011 at 20:51

fge

122k35 gold badges266 silver badges340 bronze badges

5 Comments

sasa Over a year ago

second regex returns class with method, AdminClass::display. I think that regex by F.J. can be used as second

fge Over a year ago

Well, it really can be done with one regex and no errors, but it will be a very long regex and require even more explanation ;) I can cook it up and explain it all if you want.

Andrew Clark Over a year ago

Wouldn't this also match a/b and put it all into $method?

fge Over a year ago

Yes, which is why I suggested to replace .* with more restrictive regexes. But see edit.

fge Over a year ago

Done. Now with the full pizzazz :p

Sam Sussman · Accepted Answer · 2011-12-22 20:52:19Z

-1

        $pieces = explode('/',$controller); 
        $path = '';
        for($i = $i<$pieces.length-1; $i++)
        {
          if($i != 0)
             $path+='/';
          $path += $pieces[$i];
        }
        $p2 = explode(  '::',$pieces[$pieces.length-1]);
        $class = $p2[0];
        $method = $p2[1];

edited Dec 22, 2011 at 20:52

answered Dec 22, 2011 at 20:35

Sam Sussman

1,0458 silver badges27 bronze badges

Collectives™ on Stack Overflow

Parsing string - with regex or something similar?

3 Answers 3

3 Comments

5 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

3 Comments

5 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related