Numpy array of indexes to array of elements

Question

I have a MxN numpy array containing ints, representing indexes of a big array of size K How do I convert efficiently my M*N array of indexes to an MxN array of elements?

Example :

K = ['a','b','c','d']
M = [[0,3],[2,1]]

Result :

[['a','d'],['c','b']]

Thank you!

Answer 1

We can make numpy arrays from these lists:

import numpy as np

k = np.array(K)
m = np.array(M)

and then perform a mapping with k[m] :

>>> k[m]
array([['a', 'd'],
       ['c', 'b']], dtype='<U1')

Here for every element in m , we thus "replace" it with the element stored in k at the index of the original value of m at that location.