Re^2: Grabbing numbers from a URL

[BillKSmith]

Your final comment brings up a bigger problem with the specification. If the current four-digit format allows more than four digits (only the last four of which are used), without more information, it will be impossible to distinguish between this case and the new five-digit numbers. All solutions so far assume that if there are at least five digits, all five are part of the number. I like your solution because it extends this assumption to any number of digits.

[Marshall]

I was immediately thinking that this "number" is intended to be a "unique id number". I suppose that could be a wrong assumption, but in my experience if there is some number like 12389799 and somehow those last 5 digits are "special", then the name should be "....123_89799.htm". My algorithm will work with that. If I have any influence over the file naming convention, I will put an "_" in there to separate the fields. In my opinion, stuff like a fixed 5 digit deal, maybe 00123.html is a very bad idea. Often these software things grow and maybe at some point a 6th digit is needed? Then what? In general, I like the idea of having the basic parsing being one thing and if needed the validation of those fields being a separate thing. If two digits aren't allowed, then if ($number < 100) {}. "....123_89799" and "....123_897" should be valid if the naming convention guy has his eye on the ball. These details DO matter.