Stupid ways to spell simple code

Joshua Landau joshua at landau.ws
Thu Jul 11 04:37:28 CEST 2013


On 30 June 2013 07:06, Chris Angelico <rosuav at gmail.com> wrote:
> So, here's a challenge: Come up with something really simple, and
> write an insanely complicated - yet perfectly valid - way to achieve
> the same thing. Bonus points for horribly abusing Python's clean
> syntax in the process.

This occurred to me out of the blue while working on something
related. Here's a way to remove all instances of an element from an
iterable. It's remarkably fast for it's course of action:

    from collections import deque
    from itertools import chain

    exhaust_iterable = deque(maxlen=0).extend

    def split_on(data, sentinel):
        chained = data = iter(data)
        while True:
            chunk = iter(chained.__next__, sentinel)
            yield chunk

            # Uses at least one item from chained, so "chained" and
"data" are equivilant after this.
            # This is important as "data" is raw and thus will not get
bogged down by pointless chain wrappings.
            # Yes, that is a premature optimisation. Go away.
            exhaust_iterable(chunk)

            # Throw StopIteration if not empty
            chained = chain([next(data)], data)

    def remove_all(iterable, victim):
        return list(chain.from_iterable(split_on(iterable, victim)))

    print(remove_all([1, 2, 1, 1, 0, 1, 2, 1, 1, 2, 2, 1, 0], 1))


and here it is again:


    from itertools import chain, repeat

    def remove_all(iterable, victim):
        iterable = iter(iterable)
        return list(chain.from_iterable(iter(chain([next(iterable)],
iterable).__next__, victim) for _ in repeat(...)))

    print(remove_all([1, 2, 1, 1, 0, 1, 2, 1, 1, 2, 2, 1, 0], 1))



More information about the Python-list mailing list