How to initiate 2-D array in Python

This tutorial guide will introduce different methods to initiate a 2-D array in Python. We will make a 3x5 2-D array in the following examples.

list comprehension Method

>>> column, row = 3, 5
>>> array2D = [[0 for _ in range(row)] for _ in range(column)]
>>> array2D
[[0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0]]

This nested list comprehension method creates 2-D array with the initial value as 0. Of course, you could change the initial value to any value you need to assign in your application.

Nested range Method

If you don’t care about the initial value in the 2-D array, the value 0 could be even eliminated.

In Python 2.x

>>> column, row = 3, 5
>>> A = [range(row) for _ in range(column)]
>>> A
[[0, 1, 2, 3, 4], [0, 1, 2, 3, 4], [0, 1, 2, 3, 4]]

In Python 3.x

>>> column, row = 3, 5
>>> A = [range(row) for _ in range(column)]
>>> A
[range(0, 5), range(0, 5), range(0, 5)]

We couldn’t simply use range(x) to initiate 2-D array in Python 3.x because range returns an object containing a sequence of integers in Python 3.x, but not a list of integers as in Python 2.x. range in Python 3.x in more similar to xrange in Python 2.x. range object in Python 3.x is immutable, therefore, you don’t assign items to its elements.

If you need item assignment, you need to convert the range to list object.

>>> A = [list(range(row)) for _ in range(column)]
>>> A
[[0, 1, 2, 3, 4], [0, 1, 2, 3, 4], [0, 1, 2, 3, 4]]

[0] * n Method

One Pythonic way to initiate a 2-D array could be

>>> column, row = 3, 5
>>> A = [[0]*row for _ in range(column)]
>>> A
[[0, 0, 0, 0, 0], [0, 0, 0, 0, 0], [0, 0, 0, 0, 0]]

Although we should be cautious when we use list multiplication because it simply creates a sequence with multiple times referred to one same object, we are relieved to use [0]*n here because data object 0 is immutable so that we will never encounter problems even with references to the same immutable object.

numpy Method

Besides the native Python array, numpy should be the best option to create a 2-D array, or to be more precise, a matrix.

You could create a matrix filled with zeros with numpy.zeros.

>>> import numpy as np
>>> column, row = 3, 5
>>> np.zeros(column, row)
array([[0., 0., 0., 0., 0.],
       [0., 0., 0., 0., 0.],
       [0., 0., 0., 0., 0.]])

Or initiate a matrix filled with ones with numpy.ones

>>> import numpy as np
>>> column, row = 3, 5
>>> np.ones((column, row))
array([[1., 1., 1., 1., 1.],
       [1., 1., 1., 1., 1.],
       [1., 1., 1., 1., 1.]])

You could even create a new array without initializing entries with numpy.empty

>>> import numpy as np
>>> column, row = 3, 5
>>> np.empty((5,5))
array([[6.23042070e-307, 4.67296746e-307, 1.69121096e-306,
        1.33511562e-306, 1.89146896e-307],
       [7.56571288e-307, 3.11525958e-307, 1.24610723e-306,
        1.37962320e-306, 1.29060871e-306],
       [2.22518251e-306, 1.33511969e-306, 1.78022342e-306,
        1.05700345e-307, 1.11261027e-306],
       [1.11261502e-306, 1.42410839e-306, 7.56597770e-307,
        6.23059726e-307, 1.42419530e-306],
       [7.56599128e-307, 1.78022206e-306, 8.34451503e-308,
        2.22507386e-306, 7.20705877e+159]])
Notes

It is a better solution if you want to create the empty array first and then assign the element values later. But be aware that random values are in the array so that it could be risky if you access the array by indexing before value of the corresponding index has been assigned.