Python Forum

a = np.arange(25).reshape(5, 5)
a

Output:
array([[ 0,  1,  2,  3,  4],
        [ 5,  6,  7,  8,  9],
        [10, 11, 12, 13, 14],
        [15, 16, 17, 18, 19],
        [20, 21, 22, 23, 24]])

np.diag(a, k = -1)

Output:
array([ 5, 11, 17, 23])

np.diag(a, k = -2)

Output:
array([10, 16, 22])

How to do like below:
Use feature of np.diag(), indexing specific elements then multiply 10?

a = np.full((5, 5), 1)

Output:
array([[1, 1, 1, 1, 1],
       [1, 1, 1, 1, 1],
       [1, 1, 1, 1, 1],
       [1, 1, 1, 1, 1],
       [1, 1, 1, 1, 1]])

want:

Output:
array([[1, 1, 1, 1, 1],
       [10, 1, 1, 1, 1],
       [1, 10, 1, 1, 1],
       [1, 1, 10, 1, 1],
       [1, 1, 1, 10, 1]])

Output:
array([[1, 1, 1, 1, 1],
       [1, 1, 1, 1, 1],
       [10, 1, 1, 1, 1],
       [1, 10, 1, 1, 1],
       [1, 1, 10, 1, 1]])

For the first matrix (use the same strategy for the other one)

Paul

import numpy as np
n = 5
M = np.ones( (n,n) )
i = np.arange(1,n)
M[i,i-1] = 10*M[i,i-1]

print(f'M = {M}')

Output:
M = [[ 1.  1.  1.  1.  1.]
 [10.  1.  1.  1.  1.]
 [ 1. 10.  1.  1.  1.]
 [ 1.  1. 10.  1.  1.]
 [ 1.  1.  1. 10.  1.]]