Adding length 1 dimensions with newaxis

NumPy has a nice shortcut for adding a length 1 dimension to an array. It is a little brain-bending, because it operates via array slicing:

import numpy as np

v = np.array([0, 3])
v.shape

(2,)

# Insert a new length 1 dimension at the beginning
row_v = v[np.newaxis, :]
print(row_v.shape)
row_v

(1, 2)

array([[0, 3]])

# Insert a new length 1 dimension at the end
col_v = v[:, np.newaxis]
print(col_v.shape)
col_v

(2, 1)

array([[0],
       [3]])

Read this last slicing operation as “do slicing as normal, except, before slicing, insert a length 1 dimension at the position of np.newaxis”.

In fact the name np.newaxis points to the familiar Python None object:

np.newaxis is None

True

So, you also use the np.newaxis trick like this:

row_v = v[None, :]
row_v.shape

(1, 2)