lists_vs_sets

lists_vs_sets.py

"""
Perf lists versus sets for indexed operations.
"""

import random
import time

rand = lambda i: random.randint(1, i)

now = time.time

I = 100 * 1000

# create a big list and an anloge set:
a_list = [i for i in range(I)]
a_set = {i for i in range(I)}

# we'll remove from those these (every 10th item):
difflist = [i * 10 for i in range(int(I / 10))]
diffset = {i * 10 for i in range(int(I / 10))}

t0 = now()
for i in difflist:
    if i in a_list:
        a_list.remove(i)
dtl = now() - t0
print('with lists the diff took', dtl)
remaining = len(a_list)

t0 = now()
a_set.difference_update(diffset)
assert len(a_set) == remaining  # go sure we have the same result
dts = now() - t0
print('with sets the diff took', dts)

print('=> sets are %s times faster' % int(dtl / dts))

python perf.py
with lists the diff took 9.214841842651367
with sets the diff took 0.0005748271942138672
=> sets are 16030 times faster