Michael Spencer
2/9/2008 1:07:00 AM
ki lo wrote:
> I have type variable which may have been set to 'D' or 'E'
>
> Now, which one of following statements are more efficient
>
> if type =='D' or type == 'E':
>
> or
>
> if re.search("D|E", type):
>
> Please let me know because the function is going to called 10s of
> millions of times.
>
> Thanks
> Kilo
>
You can easily find out yourself, using the timeit module. Here are some
examples to get you started:
# succeeding cases
$ python -mtimeit -s"t='D'" "if t == 'D' or t == 'E': pass"
1000000 loops, best of 3: 0.244 usec per loop
$ python -mtimeit -s"t='D'; import re" "if re.search('D|E', t): pass"
100000 loops, best of 3: 5.61 usec per loop
# failing cases
$ python -mtimeit -s"t='F'" "if t == 'D' or t == 'E': pass"
1000000 loops, best of 3: 0.399 usec per loop
$ python -mtimeit -s"t='F'; import re" "if re.search('D|E', t): pass"
100000 loops, best of 3: 4.47 usec per loop
#-> re is much worse whether or not the test succeeds
# try precompiling the re
$ python -mtimeit -s"t='D'; import re; patt = re.compile('D|E')" "if
patt.search(t): pass"
1000000 loops, best of 3: 1.93 usec per loop
#-> re still worse
# perhaps faster
$ python -mtimeit -s"t='D'" "if t in 'DE': pass"
1000000 loops, best of 3: 0.204 usec per loop
HTH,
Michael