split 1d np array into 3d array [duplicate]

Question

I have the following array:

x_train.shape
(7201,)

x_train
array([array([[0.01709922],
       [0.01946028],
       [0.01317027],
       [0.01228488],
       [0.01372365],
       [0.02148931],
       [0.01127036],
       [0.01887001],
       [0.01468282],
       [0.01235866]], dtype=float32),
       array([[0.01269068],
       [0.01193441],
       [0.01077232],
       [0.01219265],
       [0.02014277],
       [0.01250623],
       [0.01759726],
       [0.01145482],
       [0.00204748],
       [0.00372604]], dtype=float32),
       array([[0.01660118],
       [0.01931271],
       [0.02100972],
       [0.0167303 ],
       [0.02126796],
       [0.01245089],
       [0.00612399],
       [0.01128881],
       [0.01344696],
       [0.01422168]], dtype=float32),
       ...,

I want to transform it into the shape (X,10,1).

How can i do this in the first place, whats an efficient way to do this?

Do you understand why it is a 1d object dtype array in the first place? Are the arrays all the same shape? A quick count suggested that one has 11 rows. If they do match, np.stack(x_train) should work. — hpaulj
– hpaulj, Commented Aug 23, 2022 at 15:21

alani · Accepted Answer · 2022-08-23 13:19:05Z

0

Normally when you instantiate a numpy array using np.array, it will detect if the inputs comprise a regularly shaped array, and if so, it will give you a multi-dimensional array automatically.

If you have somehow managed to end up with an array of objects despite it being regularly shaped, you could convert it to a list and try again, for example:

import numpy as np

x = np.zeros(3, dtype=object)

x[0] = np.array([[1],[2],[3],[4]])
x[1] = np.array([[5],[6],[7],[8]])
x[2] = np.array([[9],[10],[11],[12]])

print(x.shape, x.dtype)  # prints (3,) object                                                                                               

y = np.array(list(x))

print(y.shape, y.dtype) )  # prints (3, 4, 1) int64

If this fails (giving you a warning about ragged sequences) and gives you back another object array, then your array was not in fact regularly shaped - in which case, check the number of elements in each of the subarrays. You will see this if you remove one of the elements in the above test code.

answered Aug 23, 2022 at 13:19

alani

13.2k3 gold badges18 silver badges34 bronze badges

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

1 Comment

cocktail Over a year ago

Thanks for your answer, however this didnt work for me. i found this answer which worked: stackoverflow.com/a/35193026/8055893

Collectives™ on Stack Overflow

split 1d np array into 3d array [duplicate]

1 Answer 1

1 Comment

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Linked

Related