Profiling in microseconds precision in Python 2.7

Question

I am on Win 10 OS, python 2.7.

Both time.time() and datetime.datetime.now() module give me millisecond precision and I would like to have in microseconds precision.

I have a python application keep running, it has its busiest time (big data flow in an out) and thus the time spent on each function would be vary. I would like to profile each function and let the application keep running, aim to identify the bottleneck of the function, where microseconds is needed in this case.

import datetime
print datetime.datetime.now()
2018-07-05 16:29:39.185000

import time
print time.time()
1530779379.18

Is it a problem of Win 10 OS? I believe there should be some lib cater this situation. Please help.

the time_ns seems to be available only on python 3.7 but 2.7 — mcsy, Jul 05 '18 at 09:00
my mistake, time.clock() should work for me. Need to try that for multiprocessing tho — mcsy, Jul 05 '18 at 09:02

score 0 · Answer 1 · answered Jul 05 '18 at 08:56

0

As written in documentation this datetime.datetime.now() actually returns you microseconds:

YYYY-MM-DD HH:MM:SS.microseconds

So in your case 185000 is microseconds

answered Jul 05 '18 at 08:56

Dilshat

1,088
10
12

I think mcsy meant that only the milliseconds part of that ever was non-zero. – maxy Jul 05 '18 at 10:13
ya anyway thanks Dilshat, I get my solution by using time.clock() – mcsy Jul 05 '18 at 10:16

Profiling in microseconds precision in Python 2.7

1 Answers1