The code above looks like it might accomplish what you asked, but I have a sneaking suspicion that you really want to do something else with that data, once you've created the large number of smaller files.

If so, you'll probably find the current design to be quite expensive, due to a lot of IO.

If I'm right, would you consider describing the purpose of this routine? We might be able to come up with a better process -- or not. ;-)

If I'm wrong, and you really just want to create a bunch of small files, I apologize for my errant hunch.