Thu, 30 Nov 2000 09:30:22 +0100
Moshe Zadka wrote:
> > BTW, one lesson to take from SETL: a vital set operation in practice is a
> > mutating "pick a 'random' element E from the set, remove it from the set,
> > and return it". An enormous number of set algorithms are of the form
> > while not S.empty():
> > pick some element from S
> > deal with it, possibly mutating S
> > This operation can't be done efficiently in Python code if the set is
> > represented by a dict (the best you can do is materialize the full list of
> > keys first, and pick one of those). That means my Set class often takes
> > quadratic time for what *should* be linear-time algorithms.
> Hmmm...actually, I've been wanting a method .key() for dictionaries
> a long time. So if we give dictionaries this one small method, then
> we *can* do this in Python.
Shouldn't be hard to do... the C API for this is already in place:
PyDict_Next(). I'd prefer a method .getitem(), though, which then
returns a tuple (key, value).
Python Pages: http://www.lemburg.com/python/