How/Why Does This Regex Work The Way It Does?

[HartleySan]

I was checking out the str_split function on php.net (http://php.net/manual/en/function.str-split.php), and the very first user comment from "atolia at gmail----- dot com" offers up a way to write your own str_split function with the preg_split function. Unfortunately, I cannot understand the regex pattern used for the function and why the regex works the way it does.

 

Specifically, the following is the alternative str_split function the user provides:

function str_split($str){
  return preg_split('//',$str);
}

 

And here is my specific question:

Why does the regex '//' split the string on every character?

 

Thank you.

[Antonio Conte]

I would guess it is because the actual expression is an empty but still a valid regular expression. (started and ended by a Char '/') As a consequence, I guess everything is a match, and it will then split on any character. While I guess this is the functional behavior, I cannot explain you why that would be.

 

The second example on preg_split() in the manual uses the same approach, btw.

[HartleySan]

Yeah, I noticed the preg_split example as well.

And yeah, I'm in the same boat as you in that I'm just assuming that when the regex is empty, it matches every character.

I suppose that's all we really can assume without knowing more about the inner-workings of regex engines.

 

I suppose it makes sense to some degree though, as regexes do typically just loop through every character in a given string. If the regex for a string is nothing, then I guess the default is to assume that every character is a match, like you said.

 

Thanks for your input, Antonio.

We'll see if Larry or someone else can provide any more details.

[Larry]

I'm not finding any documentation that explains the impact of using what seems to be an empty pattern.

[HartleySan]

Me neither, which is exactly why I asked here.

It definitely works though, so I can only assume that when an empty regex is used, every character is matched (as Antonio suggested).

 

Playing around a bit, I also noticed that unless you specify a certain flag for the preg_split command, every character (including zero-width characters) will be matched. For example, the characters represented by ^ and $ in a regex are also matched and returned.

 

But yeah, like you said, Larry, I was hoping to find some concrete information in some documentation somewhere, but no real information seems to readily exist.