numpy.diag_indices() in Python

The numpy.diag_indices() function returns indices in order to access the elements of main diagonal of a array with minimum dimension = 2. Returns indices in the form of tuple. 
to access the main diagonal of an array. 

Syntax: numpy.diag_indices(n, n_dim = 2)

Parameters : 

n : size of array, for which indices of diag elements are required along each dimension
n_dim  : [int, optional]number of dimensions. 

Return : 

Indices(as tuples) to access diagonal elements.

Code 1 :

Python3

# Python Program illustrating
# working of diag_indices()

import numpy as geek

# Creates a 5 X 5 array and returns indices of
# main diagonal elements
d = geek.diag_indices(5)
print("Indices of diagonal elements as tuple : ")
print(d, "\n")

array = geek.arange(16).reshape(4,4)
print("Initial array : \n", array)

# Here we can manipulate diagonal elements
# by accessing the diagonal elements
d = geek.diag_indices(4)
array[d] = 25
print("\n New array : \n", array)

Output :

Indices of diagonal elements as tuple : 
(array([0, 1, 2, 3, 4]), array([0, 1, 2, 3, 4])) 

Initial array : 
 [[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]
 [12 13 14 15]]

 New array : 
 [[25  1  2  3]
 [ 4 25  6  7]
 [ 8  9 25 11]
 [12 13 14 25]]

Code 2 : Manipulating 2D array 

Python

# Python Program illustrating
# working of diag_indices()

import numpy as geek

# Manipulating a 2D array
d = geek.diag_indices(3, 2)

array = geek.arange(12).reshape(4, 3)

array[d] = 111
print("Manipulated array : \n", array)

Output :

Manipulated array : 
 [[111   1   2]
 [  3 111   5]
 [  6   7 111]
 [  9  10  11]]

Code 3 : Manipulating 3D array 

Python

# Python Program illustrating
# working of diag_indices()

import numpy as geek

# Setting diagonal indices
d = geek.diag_indices(1, 2)
print("Diag indices : \n", d)

# Creating a 3D array with all ones
array = geek.ones((2, 2, 2), dtype=geek.int)
print("Initial array : \n", array)

# Manipulating a 3D array
array[d] = 0
print("New array : \n", array)

Output : 

Diag indices : 
 (array([0]), array([0]))
Initial array : 
 [[[1 1]
  [1 1]]

 [[1 1]
  [1 1]]]
New array : 
 [[[0 0]
  [1 1]]

 [[1 1]
  [1 1]]]

Note : 
These codes won’t run on online IDE’s. So please, run them on your systems to explore the working.