Fast method of converting all strings in a list to integers

Question

I've gone through the Convert all strings in a list to int post

I want to convert

results = ['1', '2', '3']

to:

results = [1, 2, 3]

I know i can do it by

1. map(int, results)

and

2. results = list(map(int, results))

I want it faster way may be using numpy or more faster.

Actual code is

from sys import stdin, stdout
import numpy as np
n = int(stdin.readline())


for i in range(0,n):
    lone = 0
    m = int(stdin.readline())
    results = stdin.readline().split()
    o = np.array(results, dtype=np.int64)
    for j in range (0,m):
        if o[j] in  o[j+1:m]:
            lone = lone +1
        elif o[j] in o [0:j]:
            lone = lone +1
        else:
            stdout.write(str(o[j]) + '\n')

    if lone == m:
        stdout.write ("-1 \n")

Please let me know if there is any methods to achieve this when trying to work with thousands of strings

Answer 1

Multiprocessing is a way to get it done faster (in addition of using Numpy):

E.g:

In [11]: from multiprocessing import Pool

In [12]: pool = Pool(10)

In [13]: pool.map(int, [str(i) for i in range(500)])

Numpy will mostly provide a memory gain as you would be dealing with primitive types instead of python objects, but will also provide a non-negligible speed gain as well. Optimizing time of an iteration like this is always done by using parallelization so I advise using both Numpy and a process pool.