Question

If I have two ndarrays:

a.shape   # returns (200,300, 3)
b.shape   # returns (200, 300)

numpy.vstack((a,b))  # Gives error

Would print out the error:
ValueError: arrays must have same number of dimensions

I tried doing vstack((a.reshape(-1,300), b) which kind of works, but the output is very weird.

Answer 1

You don’t specify what final shape you actually want. If it’s (200, 300, 4), you can use dstack instead:

>>> import numpy as np
>>> a = np.random.random((200,300,3))
>>> b = np.random.random((200,300))
>>> c = np.dstack((a,b)) 
>>> c.shape
(200, 300, 4)

Basically, when you’re stacking, the lengths have to agree in all the other axes.

[Updated based on comment:]

If you want (800, 300) you could try something like this:

>>> a = np.ones((2, 3, 3)) * np.array([1,2,3])
>>> b = np.ones((2, 3)) * 4
>>> c = np.dstack((a,b))
>>> c
array([[[ 1.,  2.,  3.,  4.],
        [ 1.,  2.,  3.,  4.],
        [ 1.,  2.,  3.,  4.]],

       [[ 1.,  2.,  3.,  4.],
        [ 1.,  2.,  3.,  4.],
        [ 1.,  2.,  3.,  4.]]])
>>> c.T.reshape(c.shape[0]*c.shape[-1], -1)
array([[ 1.,  1.,  1.],
       [ 1.,  1.,  1.],
       [ 2.,  2.,  2.],
       [ 2.,  2.,  2.],
       [ 3.,  3.,  3.],
       [ 3.,  3.,  3.],
       [ 4.,  4.,  4.],
       [ 4.,  4.,  4.]])
>>> c.T.reshape(c.shape[0]*c.shape[-1], -1).shape
(8, 3)