Lie Ryan wrote: > On 12/28/2009 8:54 PM, Peter Otten wrote: >> Francesco Bochicchio wrote: >> One with lazy chunks: >> >>>>> from itertools import chain, islice >>>>> def chunks(items, n): >> .... items = iter(items) >> .... for first in items: >> .... yield chain((first,), islice(items, n-1)) >> .... >>>>> [list(chunk) for chunk in chunks(range(10), 3)] >> [[0, 1, 2], [3, 4, 5], [6, 7, 8], [9]] > > > better: > import itertools > def chunks(items, n): > ctr = (x // n for x in itertools.count()) > return itertools.groupby(items, lambda _: next(ctr)) > > input_list = range(10) > list(list(x[1]) for x in chunks(input_list, 3)) > # [[0, 1, 2], [3, 4, 5], [6, 7, 8], [9]]
I disagree. It may be of little practical relevance that this fails after about sys.maxint items (Python 2.5) or at least slows down (Python 2.6) when Python long integer arithmetic kicks in, but I still feel uneasy about it. If I were to use groupby() I'd probably code the "counter" as ctr = cycle([False]*n + [True]*n) Peter -- http://mail.python.org/mailman/listinfo/python-list
