Multiplying each column of a d x n matrix with each column of a 1 x n vector-CodePudding

Suppose I have a 1 x 5 vector:

vect = np.array([10, 20, 30, 40, 50])

and I have a 2 x 5 matrix:

mat = np.array([
  [1, 2, 3, 4, 5],
  [6, 7, 8, 9, 10]
])

and I would like to multiply together as so: 10 * [[1], [6]] = [[10], [60]] for each column in the vector and each column in the matrix, outputting a d x n vector. How can I efficiently accomplish this with numpy? I have tried to investigate dot product, but it doesn't seem to accomplish my goals.

CodePudding user response：

vect * mat does what you want, column-wise multiplication.

CodePudding user response：

A*B should do it

import numpy as np
A = 10*(np.arange(5) 1)
B = (np.arange(10) 1).reshape([2,5])
assert (A*B == np.array([[ 10,  40,  90, 160, 250],
       [ 60, 140, 240, 360, 500]])).all()

CodePudding user response：

Following @John Zwinck's answer, if your idea is to multiply vect 1st column to mat first column and so on, vect * mat will be enough. On the other hand, if you have a second thought of using every column in vect to multiply with every column in mat, you can use numpy.tensordot.

np.tensordot(vect,mat,axes=0)

Out[15]: 
array([[[ 10,  20,  30,  40,  50],
        [ 60,  70,  80,  90, 100]],

       [[ 20,  40,  60,  80, 100],
        [120, 140, 160, 180, 200]],

       [[ 30,  60,  90, 120, 150],
        [180, 210, 240, 270, 300]],

       [[ 40,  80, 120, 160, 200],
        [240, 280, 320, 360, 400]],

       [[ 50, 100, 150, 200, 250],
        [300, 350, 400, 450, 500]]])