"stacking" Arrays In A New Dimension?

Consider, for reference: >>> x, y = np.ones((2, 2, 2)), np.zeros((2, 2, 2)) >>> np.concatenate((x, y, x, y), axis=2) array([[[ 1., 1., 0., 0., 1., 1., 0.,

Solution 1:

You can do it as follows:

>>>xx = x[..., None, :]>>>yy = y[..., None, :]>>>np.concatenate((xx, yy, xx, yy), axis=2).shape
(2, 2, 4, 2)
>>>np.concatenate((xx, yy, xx, yy), axis=2)
array([[[[ 1.,  1.],
         [ 0.,  0.],
         [ 1.,  1.],
         [ 0.,  0.]],

        [[ 1.,  1.],
         [ 0.,  0.],
         [ 1.,  1.],
         [ 0.,  0.]]],


       [[[ 1.,  1.],
         [ 0.,  0.],
         [ 1.,  1.],
         [ 0.,  0.]],

        [[ 1.,  1.],
         [ 0.,  0.],
         [ 1.,  1.],
         [ 0.,  0.]]]])
>>>

What this example does is change the shape (no data is copied) of the arrays. Slicing with None or equivalently np.newaxis adds an axis:

>>> xx.shape
(2, 2, 1, 2)
>>> xx
array([[[[ 1.,  1.]],

        [[ 1.,  1.]]],


       [[[ 1.,  1.]],

        [[ 1.,  1.]]]])
>>>