python - 从解释器和命令行使用 timeit 的时间差异

Question

从口译员那里，我得到：

>>> timeit.repeat("-".join( str(n) for n in range(10000) ) , repeat = 3, number=10000)
[1.2294530868530273, 1.2298660278320312, 1.2300069332122803] # this is seconds 

从命令行，我得到：

$ python -m timeit -n 10000 '"-".join(str(n) for n in range(10000))'
10000 loops, best of 3: 1.79 msec per loop # this is milli second 

为什么这两种情况下的时间大小有这种差异？

Answer 1

The two lines aren't measuring the same thing. In the first snippet, you're timing the calculation 0-1-2-...-9999. while in the second snippet you're timing the string concatenation "-".join(str(n) for n in range(10000)).

In addition, timeit and repeat report the total time, while the CLI averages the time over the number of iterations. So the first code actually takes 12.29 ms "per loop".