Steven Bethard wrote: > I thought it might be helpful to code some of the alternatives you've > been given and look at the timings to put things into perspective. The > code: > > -------------------- remove.py -------------------- > def remove_lc(x, lst): > lst[:] = [item for item in lst if item != x]
[snip] > $ python -m timeit -s "import remove; lst = [x % 1000 for x in > xrange(10000)]" "remove.remove_<name>(500, lst)" I do not think it measures what you think it measures. A statement in the -s option is only run once, so you have to be careful with mutable data and pass a copy of your sample to the function about to be timed. The following example illustrates the problem: $ py24 -m timeit -n5 -r1 -s"r=[0]" -s"def f(r): print r; r[0] += 1" "f(r)" [0] [1] [2] [3] [4] 5 loops, best of 1: 106 usec per loop Peter -- http://mail.python.org/mailman/listinfo/python-list
