ActiveState Code

just wrap it up. >>> timeit.Timer('a()', 'a = lambda:"foobarbaz"').timeit() 0.75941300392150879

>>> timeit.Timer('a()', 'import itertools;a = itertools.repeat("foobarbaz").next').timeit()
0.50646805763244629
>>> timeit.Timer('a()', 'import itertools;a = (lambda s:itertools.repeat(s).next)("foobarbaz")').timeit()
0.40992307662963867

Surprisingly, these are all consistent results, even the last two! I really can't explain why, but creating a function you can pass something to to do the work, rather than using itertools.repeat().next yourself is actually faster on the actual calls!

Meaningful timing results. The timeit module can be tricky to use. Try specifying exactly how many repetitions you want and run the test multiple times in the a single script. Try to get your system as quiet as possible (switch-off other processes and net activity). Then, you with have a fighting chance of getting results that make sense:

from timeit import Timer

setup = '''
import itertools
a = lambda: 'foobarbaz'
b = itertools.repeat('foobarbaz').next
c = (lambda s:itertools.repeat(s).next)("foobarbaz")
'''

for i in (1,2,3,4,5):
    for stmt in 'a() b() c()'.split():
        print stmt, min(Timer(stmt, setup).repeat(7, 1000000))
    print