CODEWITHSUNDEEP
pythonarraysnumpy
comalex3
comalex3

Reputation: 2606

how to swap columns of the matrix, numpy

You need to sort the columns by decay values ​​of the diagonal elements, 0.884>0.749>0.640, this swap 1 and 3 column

numpy.array(
    [
        [ 0.640 -0.655  0.399]
        [ 0.617  0.749  0.239]
        [-0.456  0.093  0.884]
     ]

to receive the result :

numpy.array(
    [
        [ 0.399 -0.655  0.640]
        [ 0.239  0.749  0.617]
        [-0.884  0.093 -0.456]
     ]

Upvotes: 0

Views: 1827

Answers (3)

Geeklhem
Geeklhem

Reputation: 689

I would do :

a[: , numpy.argsort(a.diagonal())[::-1] ]
  • a.diagonal to get the diagonal values with [::-1] to get them in reverse order
  • numpy.argsort to get the new order of the columns

Upvotes: 1

Harpal
Harpal

Reputation: 12587

You could use advanced splicing:

>>> import numpy as np
>>> a = np.arange(25).reshape(5,5)
>>> a
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]])
>>> a[:,[0,4]] = a[:,[4,0]]
>>> a
array([[ 4,  1,  2,  3,  0],
       [ 9,  6,  7,  8,  5],
       [14, 11, 12, 13, 10],
       [19, 16, 17, 18, 15],
       [24, 21, 22, 23, 20]])
>>>

Upvotes: 0

Akavall
Akavall

Reputation: 86188

I think this is what you are looking for:

>>> a
array([[ 0.64 , -0.655,  0.399],
       [ 0.617,  0.749,  0.239],
       [-0.456,  0.093,  0.884]])
>>> a[:, np.argsort(a.diagonal() * -1)]
array([[ 0.399, -0.655,  0.64 ],
       [ 0.239,  0.749,  0.617],
       [ 0.884,  0.093, -0.456]])

Upvotes: 2

Related Questions