[Python-ideas] data structures should have an .any() method

[Gerald Britton]

Interesting that your return(next()) example runs about 60% slower
than the "for x in l: return x" example.  Must be the function call
overhead.

On Fri, Sep 4, 2009 at 3:36 PM, Stefan Behnel<stefan_ml at behnel.de> wrote:
> Guido van Rossum wrote:
>> On Fri, Sep 4, 2009 at 2:35 AM, Stefan Behnel wrote:
>>> I just had a discussion with a co-worker, and we noticed that there are use
>>> cases where you just want the only element in a data structure, or just any
>>> of the elements in a data structure because you know that they all contain
>>> the same information (with respect to what you are looking for, at least).
>>>
>>> If you want all items, you can iterate, but if you just want any item or
>>> the only item, it's inefficient (and not very explicit code) to create an
>>> iterator and take the element out.
>>
>> I assure you it's not slow.
>
> Not in absolute numbers, but certainly slower than necessary:
>
> $ python2.6 -m timeit -s 'l=[1]' 'l[0]'
> 10000000 loops, best of 3: 0.0977 usec per loop
>
> $ python2.6 -m timeit -s 'l=[1]' 'next(iter(l))'
> 1000000 loops, best of 3: 0.523 usec per loop
>
>
>> next(iter(x)) is probably as good as it
>> gets -- I don't think we need another way to say that in fewer words.
>
> I'm fine with such a decision, given that it's trivial to wrap this into
> your own function. That doesn't make it much faster:
>
>
> $ python2.6 -m timeit -s 'l=[1]' -s 'def getany(l):'
>                                 -s '  return l[0]'
>                      'getany(l)'
> 1000000 loops, best of 3: 0.34 usec per loop
>
> $ python2.6 -m timeit -s 'l=[1]' -s 'def getany(l):' \
>                                 -s '  for x in l:' \
>                                 -s '    return x' \
>                      'getany(l)'
> 1000000 loops, best of 3: 0.454 usec per loop
>
> $ python2.6 -m timeit -s 'l=[1]' -s 'def getany(l):' \
>                                 -s '  return next(iter(l))' \
>                      'getany(l)'
> 1000000 loops, best of 3: 0.743 usec per loop
>
>
> but, admittedly, that's still not slow in absolute numbers.
>
> Stefan
>
> _______________________________________________
> Python-ideas mailing list
> Python-ideas at python.org
> http://mail.python.org/mailman/listinfo/python-ideas
>



-- 
Gerald Britton