Comment by driggs | Hacker News Reader

Comment by driggs | Hacker News Reader

Funny that Hack #1 compares list versus set value lookup, but the timer doesn't include the time to copy the list into a set. Hack #2 warns against unnecessary copying, and the time for copying the list is almost the same as the performance gain in Hack #1.

In fact, creating a set takes longer than copying a list since it requires hash insertion, so it's actually much faster to do the opposite of what they suggest for #1 (in the case of a single lookup, for this test case).

Here's the results with `big_set = set(big_list)` inside the timing block for the set case:

    List lookup: 0.013985s
    Set lookup:  0.052468s

I'm surprised that this was not the first complaint. It's pretty clear "Hack" #1 isn't going to work.

  import random
  import time

  def timeit(func, _list, n=1000):
      start = time.time()
      for _ in range(n):
          func(_list=_list )
      end = time.time()
      print(f"Took {end-start} s")
      return

  def lstsearch(_list):
      sf = random.randint(0,len(_list))
      if sf in _list:
          return
      return

  def setsearch(_list):
      sf = random.randint(0, len(_list))
      if sf in set(_list):
          return
      return

  mylist = list(range(100000))

  timeit(lstsearch, mylist)
  timeit(setsearch, mylist)

  ----
  Took 0.23349690437316895 s
  Took 0.8901607990264893 s

Reply on news.ycombinator.com