On Jan 5, 4:14 pm, [EMAIL PROTECTED] wrote:
> On Jan 5, 5:07 pm, [EMAIL PROTECTED] wrote:
>
> > Hello, Paul and Arnaud.
> > While I think about your answers: do you think there is any way to
> > avoid shuffle?
> > It may take unnecessary long on a long list most of whose elements
> > have the property.
You could generate a random shuffle of range(len(seq)) lazily, and use
that to iterate over your sequence.
import random
import itertools
def randxrange(n):
"Generate the numbers 0 to n-1 in a random order"
x = range(n)
for i in xrange(n):
r = random.randrange(n - i)
yield x[r]
x[r] = x[n - i - 1]
def shuffled(seq):
"Generate the elements of seq in a random order"
return (seq[i] for i in randxrange(len(seq)))
def pick_random(seq, prop):
return itertools.ifilter(prop, shuffled(seq)).next()
--
Paul Hankin
--
http://mail.python.org/mailman/listinfo/python-list
