[docs] Better roundrobin itertools recipe, IMHO

[David Lindquist]

Greetings!

I recently perused the documentation for the itertools module (
http://docs.python.org/2/library/itertools.html), and attempted to
understand the examples in the Recipes section.

The roundrobin example was particularly vexing me, so I set out to code up
an alternative that was more clear and concise. I believe I succeeded and
wondered if my code might be a better example for the documentation (with
due respect to George Sakkis!).

Not only is it fewer lines, it is more straightforward and idiomatic, I
think. Plus, it's more than twice as fast to boot (on my machine at least,
according to timeit):

from itertools import *

def roundrobin(*iterables):
    "roundrobin('ABC', 'D', 'EF') --> A D E B F C"
    sentinel = object()
    it = chain.from_iterable(izip_longest(fillvalue=sentinel, *iterables))
    return (i for i in it if i is not sentinel)

if __name__ == '__main__':
    import timeit
    print(timeit.timeit("list(roundrobin('ABC', 'D', 'EF'))",
                        setup="from __main__ import roundrobin"))

A typical result from timeit:

$ python roundrobin_mine.py
5.45610809326
$ python roundrobin_from_docs.py
13.8604800701

What do you think? Might this version supplant the current one in the
documentation?

Cheers,

David Lindquist
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/docs/attachments/20140220/43e2fc82/attachment.html>