Python反向词性能

Question

I was trying to reverse words in Python. Simple problem (leetcode).

Solution 1 : (20 ms - 99.54% percentile)

words = reversed([word.strip() for word in words if word != ''])

answer = " ".join(words).strip()

Solution 2 : (24 ms - 48% percentile)

words = [word.strip() for word in words if word != '']

answer = " ".join(reversed(words)).strip()

I am trying to find the reasoning for this variance?

Answer 1

Run a benchmark using timeit ( https://docs.python.org/3.7/library/timeit.html ) to see that the time is more or less the same (for my guessed input sample):

def opt1(words):
  words = reversed([word.strip() for word in words if word != ''])
  answer = " ".join(words).strip()
  return answer

def opt2(words):
  words = [word.strip() for word in words if word != '']
  answer = " ".join(reversed(words)).strip()
  return answer

import timeit, functools

t_opt1 = timeit.Timer(functools.partial(opt1, words))
t_opt2 = timeit.Timer(functools.partial(opt2, words))
# run 5000 times each method:
print('opt1: ', t_opt1.timeit(5000))
print('opt2: ', t_opt2.timeit(5000))

Answer 2

The only difference here is in when references to the temporary list are discarded (during join in the first case, not until function exit in the second case). But that's unlikely to matter unless the timing framework is broken in some way (it should incorporate any clean up times in both cases).

Otherwise, barring some alternate interpreter with weird behaviors being involved, both sets of code should behave identically (the op codes are identical, aside from swapping the order depending on whether the list is stored first or reversed is called first); timing differences seem more likely to be random than anything else.