NumPy eye()

import numpy as np

# create a 3x3 array with 1s in diagonal and 0s elsewhere
array1 = np.eye(3)

print(array1)

'''
Output:
[[1. 0. 0.]
 [0. 1. 0.]
 [0. 0. 1.]]
 '''

eye() Syntax

The syntax of eye() is:

numpy.eye(N, M = None, k = 0, dtype = float, order = 'C')

Note: Here, M = None implies that M = N by default.

eye() Arguments

The eye() method takes the following arguments:

• N- Number of rows in the output (can be int)
• M (optional)- Number of columns in the output (can be int)
• k (optional) - the diagonal in question(can be int)
• dtype (optional) - the datatype of the resultant array
• order (optional) - the memory layout of the array (can be 'C' or 'F')

eye() Return Value

The eye() method returns the resultant array with shape NxM with all elements as 0 except the kth diagonal where all elements are 1.

Example 1: Create a Square Array Using eye()

import numpy as np

# create a square 2x2 array with 1s at diagonal
array1 = np.eye(2)
print('2x2 square array\n', array1)

# create a square 2x2 array with 1 above diagonal
array2 = np.eye(2, k=1)
print('2x2 square array with k = 1\n', array2)

# create a square 2x2 array with 1 below diagonal
array3 = np.eye(2, k=-1)
print('2x2 square array with k = -1\n', array3)

Output

2x2 square array
 [[1. 0.]
 [0. 1.]]
2x2 square array with k = 1
 [[0. 1.]
 [0. 0.]]
2x2 square array with k = -1
 [[0. 0.]
 [1. 0.]]

Example 2: Create a Non-Square Array Using eye()

If we pass both N and M as arguments, a non-square array is created.

Let us look at an example.

import numpy as np

# create a  2x3 array with 1s at diagonal
array1 = np.eye(2, 3)
print('2x3  array\n', array1)

# create a  3x2 array with 1s at diagonal
array2 = np.eye(3, 2)
print('3x2  array\n', array2)

Output

2x3  array
[[1. 0. 0.]
 [0. 1. 0.]]
3x2  array
 [[1. 0.]
 [0. 1.]
 [0. 0.]]

Example 3: Using the Optional dtype Argument With eye()

The optional dtype argument specifies the data type of the created array.

Let's look at an example.

import numpy as np

# create a  2x2 float array 
array1 = np.eye(2, )

print('2x2 float  array\n', array1)

# create a  2x2 integer array 
array2 = np.eye(2, dtype = int)

print('2x2 int  array\n', array2)

# create a  2x2 string array 
array3 = np.eye(2, dtype = str)

print('2x2 string  array\n', array3)

Output

2x2 float  array
[[1. 0.]
 [0. 1.]]
2x2 int  array
 [[1 0]
 [0 1]]
2x2 string  array
 [['1' '']
 ['' '1']]

Note: When the array is converted to a string array, the 0s are replaced by empty strings.

Using the Optional order Argument With eye

With eye(), there are two possible options for order argument.

• 'C': C stands for C style order where the array is filled row-wise.
• 'F': F stands for Fortran-style order where the array is filled column-wise.