Time various methods of removing a possibly-present item from a dict

time_del.py

#!/usr/bin/python

import time

def new_d():
    return {
        1: 2, 3: 4, 5: 6, 7: 8, 9: 10,
        11: 12, 13: 14, 15: 16, 17: 18, 19: 20
    }

def time_func(iterations, f, *args):
    start = time.time()
    for i in xrange(iterations):
        f(*args)
    return time.time() - start

def try_del(key):
    d = new_d()
    try:
        del d[key]
    except KeyError:
        pass

def if_del(key):
    d = new_d()
    if key in d:
        del d[key]

def pop_del(key):
    d = new_d()
    d.pop(key, None)

def succeed(f):
    f(3)

def fail(f):
    f(4)

iterations = 1000000
print "pop succeed:", time_func(iterations, succeed, pop_del)
print "pop fail:", time_func(iterations, fail, pop_del)
print "try succeed:", time_func(iterations, succeed, try_del)
print "try fail:", time_func(iterations, fail, try_del)
print "if succeed:", time_func(iterations, succeed, if_del)
print "if fail:", time_func(iterations, fail, if_del)

Well, I didn't go into nearly the same level of detail, but I did feel like putting the results into a table allowed for easier consideration of what the data could tell us.

I should probably note that the level of precision in these tables is arbitrary, and we can all agree that a single run on my random pc isn't sufficient to warrant these precision levels. I chose to round to these points in case anyone else had a larger data set to compare this small sample against.

Table 1: execution time (in milliseconds) rounded to to the nearest microsecond

case	succeed	fail
pop	45.673 ms	49.161 ms
try	56.651 ms	72.481 ms
if	58.166 ms	42.423 ms

Table 2: execution times relative to the slowest test case

case	succeed	fail
pop	63.014%	67.8%
try	78.16%	100.0%
if	80.25%	58.5%

Table 2 data has been rounded to the nearest tenth of a percent (that's about +-70 microseconds for the data from my pc)