Vectorisation - applying the function to each element of the matrix

Question

I have two arrays and the function

nparr1 = np.array([1,2])
nparr2 = np.array([[0.4, 0.8, 1.2, 1.6, 2]])

def fun(a,b):
    return a + b

I would like my output to be matrix

result = [
    [fun(1,0.4), fun(1,0.8), fun(1,1.2), fun(1,1.6), fun(1,2)],
    [fun(2,0.4), fun(2,0.8), fun(2,1.2), fun(2,1.6), fun(2,2)]
]

I figured out that running

np.array(np.meshgrid(nparr1, nparr2)).T

results in intermediate step of building this matrix

intermidiate_matrix = [[[1.  0.4]
  [1.  0.8]
  [1.  1.2]
  [1.  1.6]
  [1.  2. ]]

 [[2.  0.4]
  [2.  0.8]
  [2.  1.2]
  [2.  1.6]
  [2.  2. ]]]

It looks like a step in the right direction. But I am not sure how to proceed from here. Summation in the fun is arbitrary and used only as an example. The key things is that function takes two params and reduces it to single value. I was looking into np.vectorize but didn't have much sucess.

Answer 1

What you want to do is a common operation in numpy and it makes sense to make use of broadcasting. This is a central concept in numpy and I really recommend reading that page in the documentation and learning in detail how that works if you plan to use numpy.