This is an example of something which someone else somewhere in the world already wrote, tested, debugged, and left lying around on the internet. Re-using other people's source code is a good thing, in almost every case. So long as the source code is good quality.

"Reinventing the wheel", you've probably heard that term before; a lot of programmers reinvent the wheel on a nearly hourly basis. Your code only handles one specific kind of path, which is an artificial and fairly severe limitation. If you found a standard library for parsing filenames, you could handle any sort of path you want. Perl, for example, has File::Basename, which can handle filenames from Unix, Windows, AIX, VMS, and AmigaOS.

Another good thing about code reuse is that you don't have to care about that code. If there's a bug in the filename parsing code, it's not your problem. You're doing fileparse($my_filename), and that's it. You update your File::Basename module (or whatever) and your one line of code still works without having to change anything at all.