On Jan 4, 2008 3:47 PM, Neil Cerutti <[EMAIL PROTECTED]> wrote:
> On Jan 4, 2008 2:55 PM, <[EMAIL PROTECTED]> wrote:
>
> > Hello,
> > This is a question for the best method (in terms of performance
> > only) to choose a random element from a list among those that satisfy
> > a certain property.
>
>
> > A simple approach is:
> >
> > import random
> > def random_pick(a_list,property):
> > '''Returns a random element from a list that has the property
> >
> > Returns None if no element has the property
> > '''
> > random.shuffle(a_list)
> > for i in a_list:
> > if property(i): return i
>
>
> I'm pretty sure you don't want to use a destructive random_pick function.
> You'll have to shuffle a copy instead to avoid that problem.
>
I thought of another one based on combining the above with the linear search
idea, minimizing the calls to the predicate function.
indexes = range(len(a_list))
random.shuffle(indexes)
for ix in indexes:
if predicate(a_list[ix])
return a_list[ix]
raise ValueError('no matching element in list')
--
Neil Cerutti
--
http://mail.python.org/mailman/listinfo/python-list
