I have been extensively using numpy arrays recently, and have discovered that what I used to do with meshgrids could be done with smaller 1D arrays:

import numpy as np

def myfun(x,y):
    return np.sin(x**2 + y**2) / (x**2 + y**2)

# create 2 arrays, each 1D but not on the same dimension
x = np.array([1,2,3,4,5]).reshape(1,-1)
y = np.array([6,7,8]).reshape(-1,1)

# make some computation
z = myfun(x,y)

# same with full 2D arrays
xx, yy = np.meshgrid(x,y,indexing='xy')
zz = myfun(xx,yy)


if (zz == z).all():
    print("They are the same array!")

So when should I use the full (Ny,Nx) xx and yy instead of the 1D versions?