How to revert a vertical numpy array (1D) to its usual horizontal default form?

Question

Ok, it has been asked many times how to convert a 1D numpy array to a vertical one. The most used option is, if

In [2]: a = np.array([1, 2, 3, 4])
In [3]: a
Out[3]: array([1, 2, 3, 4])

we usually do:

In [5]: a = a[ : , np.newaxis]
In [6]: a
Out[6]:
array([[1],
     [2],
     [3],
     [4]])

because we know that a.T doesn't work (something that people with a more math background that CS, like me, find a little bit shocking...).

My question is, if you receive an array (1D), that is already vertical, how do you transform it to horizontal?

Answer 1

There is no such thing as a horizontal or vertical 1D array in numpy. A 1D array is just a 1D array. It is OK if you want to call a 2D array with shape, (4, 1), a "vertical" array. Then a 2D array with shape, (1, 4), would be a "horizontal" array. In that case the transpose will work as you expect it should.

The transpose a[:, np.newaxis].T gives a horizontal array with shape, (1, 4). You can always check the shape of arrays with print(a.shape). If you want to go back to a 1D array, you can call a.squeeze(), which will return an array with shape, (4,).

import numpy as np

a = np.array([1,2,3,4])
print("Array {0} has shape {1}.".format(a, a.shape))

print(a[:, None].shape)
print((a[:, None].T).shape)
print(a[:, None].squeeze().shape)

Returns,

Array [1 2 3 4] has shape (4,).
(4, 1)
(1, 4)
(4,)