Advantage of itertools over range

fastai fastai dev
1

I noticed that in the Data Loader class, get_idxs is written as follows

def get_idxs(self):
        idxs = Inf.count if self.indexed else Inf.nones
        if self.n is not None: idxs = list(itertools.islice(idxs, self.n))
        if self.shuffle: idxs = self.shuffle_fn(idxs)
        return idxs

What is the advantage of creating an infinite count of indexes and slicing them instead of using something like range?

2

self.n might be None.

2 Likes
3

If self.n is None it means something is broken, correct? In this case, we return an infinite list of nones. But when I print that, it shows me an empty list. How does that work?

Screen Shot 2020-03-18 at 2.06.38 PM
Screen Shot 2020-03-18 at 2.06.38 PM2226×408 26.9 KB

Also if you could point me to some good resources to learn about the functional programming stuff that you mention, that’ll be great.