"John Roth" <[EMAIL PROTECTED]> wrote: > The other is as the module itself. Let's take a simple > example. Assume you have a directory named breakfast > which contains modules named spam.py, eggs.py, > toast.py and jam.py, and that the directory containing > breakfast is on the PYTHONPATH. > > If it try to import spam.py by writing > > import breakfast.spam > > it won't work because the breakfast directory > doesn't contain an __init__.py file. > > If I then add __init__.py to the breakfast directory, > the import will work, and the result will be *two* > modules loaded. The first module will be bound to > the identifier 'breakfast' in your module. It will be > completely empty except for the identifier 'spam' > which will have the spam module bound to it.
There is also an alternative form of import that does not load the intermediate modules (actually packages), only the last one: >>> import breakfast.spam as spam Here you can refer to names defined in breakfast.spam as spam.eat(), spam.slice, etc. However you can't refer to breakfast; if you do, you'll get a NameError exception: >>> import breakfast.spam as spam >>> breakfast.spam NameError: name 'breakfast' is not defined In either form of import though, the __init__.py has to be in the breakfast directory. HTH, George -- http://mail.python.org/mailman/listinfo/python-list
