how to multiply 2 numpy array with different dimensions

Question

I try to multiply 2 matrix x,y with shape (41) and (41,6) as it is supposed to broadcast the single matrix to every arrow in the multi-dimensions

I want to do it as :

x*y

but i get this error

ValueError: operands could not be broadcast together with shapes (41,6) (41,) 

Is there anything I miss here to make that possible ?

Answer 1

Broadcasting involves 2 steps

• give all arrays the same number of dimensions

• expand the 1 dimensions to match the other arrays

With your inputs

(41,6) (41,)

one is 2d, the other 1d; broadcasting can change the 1d to (1, 41), but it does not automatically expand in the other direction (41,1).

(41,6) (1,41) 

Neither (41,41) or (6,41) matches the other.

So you need to change your y to (41,1) or the x to (6,41)

x.T*y
x*y[:,None]

I'm assuming, of course, that you want element by element multiplication, not the np.dot matrix product.