Numerical operations using numpy

Creating numpy array:

Input

import numpy as np
my_array = np.array([1, 3, 4])
print(my_array)

Output

[1 3 4]

Input

type(my_array)

Output

numpy.ndarray

Create a two-dimensional array/matrix:

Input

my_array2 = np.array([[1, 2, 4], [3, 1, 5]])
print(my_array2)

Output

[[1 2 4]
 [3 1 5]]

Check dimension of an array:

Input

my_array2.ndim

Output

2

Input

my_array2.shape

Output

(2, 3)

Input

my_array2.size

Output

6

Input

my_array2.dtype

Output

dtype('int64')

We can specify the dtype explicitly if you need to.

Input

my_array3 = np.array([[1, 4, 7], [1, 9, 32]], dtype=np.float64)
my_array3.dtype

Output

dtype('float64')

We can switch type of an existing array:

Input

my_array3.astype(np.int64)

Output

array([[ 1,  4,  7],
       [ 1,  9, 32]])

Alternative way:

Input

my_array3 = np.array(my_array3, dtype=np.float64)
print(my_array3)

Output

[[ 1.  4.  7.]
 [ 1.  9. 32.]]

Creating zero matrices:

Input

my_zero_array = np.zeros((2, 3))
print(my_zero_array)

Output

[[0. 0. 0.]
 [0. 0. 0.]]

Matrix of ones:

Input

my_ones_array = np.ones((2, 3))
print(my_ones_array)

Output

[[1. 1. 1.]
 [1. 1. 1.]]

If we need to fill the matrix with some other constant, we can first create ones_matrix and multiply. However there is another function to do the same.

Input

my_const_array = np.full([2, 3], 5)
print(my_const_array)

Output

[[5 5 5]
 [5 5 5]]

Create arrays with incremental sequences:

Input

my_array4 = np.arange(1, 10, 2)
print(my_array4)

Output

[1 3 5 7 9]

Input

my_array5 = np.linspace(1, 10, 10)
print(my_array5)

Output

[ 1.  2.  3.  4.  5.  6.  7.  8.  9. 10.]

Create meshgrid:

Input

x = np.array([-1, 0, 1])
y = np.array([-2, 0, 2])

X, Y = np.meshgrid(x, y)
print(X, '\n\n', Y)

Output

[[-1  0  1]
 [-1  0  1]
 [-1  0  1]]

 [[-2 -2 -2]
 [ 0  0  0]
 [ 2  2  2]]

Other useful matrices:

Input

np.identity(4)

Output

array([[1., 0., 0., 0.],
       [0., 1., 0., 0.],
       [0., 0., 1., 0.],
       [0., 0., 0., 1.]])

Input

np.eye(4)

Output

array([[1., 0., 0., 0.],
       [0., 1., 0., 0.],
       [0., 0., 1., 0.],
       [0., 0., 0., 1.]])

Input

np.eye(4, k=1)

Output

array([[0., 1., 0., 0.],
       [0., 0., 1., 0.],
       [0., 0., 0., 1.],
       [0., 0., 0., 0.]])

Input

np.eye(4, k=-1)

Output

array([[0., 0., 0., 0.],
       [1., 0., 0., 0.],
       [0., 1., 0., 0.],
       [0., 0., 1., 0.]])

Input

np.diag(np.arange(0, 20, 5))

Output

array([[ 0,  0,  0,  0],
       [ 0,  5,  0,  0],
       [ 0,  0, 10,  0],
       [ 0,  0,  0, 15]])

Create random matrix:

Input

np.random.rand(3)

Output

array([0.49067738, 0.79451961, 0.87339187])

Input

np.random.rand(2, 3)

Output

array([[0.858099  , 0.86235168, 0.48305685],
       [0.23118357, 0.87312982, 0.1940535 ]])

Normally distributed random numbers:

Input

my_array_10 = np.random.randn(5); my_array_10

Output

array([ 0.55956033, -0.37623696,  1.78995147,  0.05591339, -0.43796257])

Expand matrix with numpy.tile():

Input

my_array_11 = np.array([[1, 3], [2, 5]]); my_array_11

Output

array([[1, 3],
       [2, 5]])

Input

my_array_12 = np.tile(my_array_11, 2); my_array_12

Output

array([[1, 3, 1, 3],
       [2, 5, 2, 5]])

Input

my_array_12 = np.tile(my_array_11, (3, 2)); my_array_12

Output

array([[1, 3, 1, 3],
       [2, 5, 2, 5],
       [1, 3, 1, 3],
       [2, 5, 2, 5],
       [1, 3, 1, 3],
       [2, 5, 2, 5]])

Indexing and slicing matrix:

Input

my_array = np.arange(5); my_array

Output

array([0, 1, 2, 3, 4])

Input

print(my_array[0], my_array[-1])

Output

0 4

Reversing an array:

Input

my_array[::-1]

Output

array([4, 3, 2, 1, 0])

Indexing in higher dimensional array:

Input

my_array = np.array([[1, 3, 2], [1, 4, 2], [5, 3, 1]]); print(my_array)

Output

[[1 3 2]
 [1 4 2]
 [5 3 1]]

Input

my_array[2, 1]

Output

3

Input

my_array[2, :]

Output

array([5, 3, 1])

Masking/substituting array values based on criteria:

import numpy as np

a = np.random.rand(10)
print(a)

# sample output
[0.80182435 0.94238444 0.42435805 0.03047743 0.20772764 0.63607548
 0.17445815 0.32827581 0.5017025  0.33546651]

b = np.array(a)  # make a deep copy
b[a>0.5] = 1
b[a<=0.5] = 0

print(b)

# output
[1. 1. 0. 0. 0. 1. 0. 0. 1. 0.]