Question

I have two matrixes

A = array([[ 12.,   0.],[  0.,   4.]])
B = array([[ 13.,   5.],[  -1.,   -5.]])

and I want to get a third one whose elements correspond to the maximum of the previous matrixes. For instance I would like to produce something like

C = array([[ 13.,   5.],[  0.,   4.]])

Is there any vectorial operation I could do to make the result faster?

Answer 1

It is easier to use numpy array instead of arrays. With a numpy array you have the np.where function to solve this:

    import numpy as np
    A = np.array([[ 12.,   0.],[  0.,   4.]])
    B = np.array([[ 13.,   5.],[  -1.,   -5.]])
    C = np.where(A>B,A,B)
    >>> C
    array([[ 13.,   5.],
    [  0.,   4.]])

This works like : np.where(condition, [returnvalue if true, returnvalue if false])
If you don’t pass the optional return parameters, you will get an array with the indexes where the condition is true.