numpy: summing every element of numpy array with every element of another

Question

I'm coming to python from Matlab. In Matlab, given two vectors that are not necessarily the same length, they can be added if one is a row vector and one is a column vector.

v1 = [1 3 5 7]

v2 = [2 4 6]'

v1 + v2

ans =

 3     5     7     9
 5     7     9    11
 7     9    11    13

I am trying to produce the same behavior in python given two numpy arrays. Looping first came to mind:

import numpy as np
v1 = np.array([1,3,5,7])
v2 = np.array([2,4,6])
v3 = np.empty((3,4,))
v3[:] = np.nan

for i in range(0,3):
    v3[i,:] = v1 + v2[i]

Is there a more concise and efficient way?

In your MATLAB size(v1) is (1,4), size(v2) is (3,1). Just make the numpy arrays the same shape. (though the numpy v1 is ok with (4,) shape). — hpaulj
– hpaulj, Commented Aug 6, 2020 at 15:39

jkr · Accepted Answer · 2020-08-06 15:29:23Z

1

import numpy as np

v1 = np.array([1, 3, 5, 7])
v2 = np.array([2, 4, 6])

v1 + v2[:, None]

You can read more about numpy's broadcasting rules.

answered Aug 6, 2020 at 15:29

jkr

19.6k5 gold badges48 silver badges78 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Ashish · Accepted Answer · 2020-08-06 15:45:14Z

0

Try this out:

for i in v2:
  z = []
  for j in v1:
    z.append(i+j)
  
  print(z)

answered Aug 6, 2020 at 15:45

Ashish

1

1 Comment

jkr Over a year ago

You typically don't want to use for loops with numpy arrays. Vectorized approaches are often faster and cleaner.

Collectives™ on Stack Overflow

numpy: summing every element of numpy array with every element of another

2 Answers 2

Comments

1 Comment

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

1 Comment

Linked

Related