Python Forum

Hello everyone,
I hope you are well.

I wonder which of those two methods for dealing with arrays would be faster in python:
method 1: define array at the beginning of the code as np.zeros for example, then fill the elements x[1] , x[2] ... etc, within the code.
method 2: define array at the beginning of the code as empty array: x = [] , and then fill the elements x[1] , x[2] ... etc, within the code.

Thanks

You should time it.

import timeit

test_setup = "import numpy as np"
test_code = '''
x = np.zeros(100000)
for i in range(len(x)):
    x[i] = i
'''
print('np.zeros', timeit.timeit(stmt=test_code, setup=test_setup, number=1000))

test_setup = ""
test_code = '''
x = []
for i in range(100000):
    x.append(i)
'''
print('append', timeit.timeit(stmt=test_code, setup=test_setup, number=1000))

Scenario 1:
You make a list and fill it with data. Task done.

Scenario 2:
You've to repeat this one million times.

If scenario 2 is the case, then preallocate the numpy.array once.
Then you can work with the same list one million times without creating new lists/arrays.
Of course, this depends on the algorithm you're using.

Allocating memory is slow.