How to convert array of numpy objects to an array containing the elements of every object?

Question

I have an numpy objects array with shape (60000,) and each of the 60000 elements is a (32,32,3)array. My question is how to convert the (60000,) array to a (32,32,3,60000) array.

• BallpointBen: docs.scipy.org/doc/numpy/reference/generated/…
• Wang Liang: you means (6000, 32, 32, 3) -> (32, 32, 3, 6000)?
• gd13: Not exactly, the way data was generated the array is a numpy object array. data.shape prints (6000,) and data[1].shape print (32,32,3). I want data.shape to print (60000,32,32,3).
• hpaulj: Try np.stack.
• gd13: concatenate and reshape worked

Answer 1

import numpy as np
class Obj():
 def __init__(self,i):
 self.i = i
l = np.array([np.array([Obj(i) for i in range(64*3)]).reshape(8,8,3)
 for _ in range(100)])
print(l.shape)
#Output: (100, 8, 8, 3)
print(np.transpose(l,(1,2,3,0)).shape)
#Output: (8, 8, 3, 100)