Issue 22446: Shortening code in abc.py (original) (raw)

Created on 2014-09-19 20:30 by cool-RR, last changed 2022-04-11 14:58 by admin. This issue is now closed.

Messages (5)
msg227117 - (view)	Author: Ram Rachum (cool-RR) *	Date: 2014-09-19 20:30
Can't this code: class Sequence(Sized, Iterable, Container): # ... def __contains__(self, value): for v in self: if v == value: return True return False Be shortened into this: class Sequence(Sized, Iterable, Container): # ... def __contains__(self, value): return any(item == value for value in self) Which can even fit on one line with a lambda: class Sequence(Sized, Iterable, Container): # ... __contains__ = lambda self: any(item == value for value in self)
msg227120 - (view)	Author: Serhiy Storchaka (serhiy.storchaka) *	Date: 2014-09-19 20:57
This is slower. >>> import timeit >>> class A(list): ... def __contains__(self, value): ... for v in self: ... if v == value: ... return True ... return False ... >>> timeit.timeit('500 in x', setup='from __main__ import A; x = A(range(1000))', number=10000) 1.1222619999971357 >>> class B(list): ... def __contains__(self, value): ... return any(v == value for v in self) ... >>> timeit.timeit('500 in x', setup='from __main__ import B; x = B(range(1000))', number=10000) 2.05952100000286
msg227121 - (view)	Author: Ram Rachum (cool-RR) *	Date: 2014-09-19 20:59
Oh. I wonder why `any` is slow like that, you'd figure it's be optimized.
msg227124 - (view)	Author: Serhiy Storchaka (serhiy.storchaka) *	Date: 2014-09-19 21:15
Because in first case there is one iterator, iter(self), and in second case there are two iterators: iter(self) and iter((v == value for v in self)).
msg227125 - (view)	Author: Ram Rachum (cool-RR) *	Date: 2014-09-19 21:21
Thanks for the clarification. Oh well, sad to see the more verbose code win, but I guess that's life. I tried on PyPy but the difference was even more pronounced, 0.008922450399566156 for the long version and 0.042124665810088044 for the short version.