I am wondering how to create a numpy array in a numpy array. Each array in the main array should be able to have a different length. The main array should be the index number to the array object inside.
I already have an example of this type of array but I want to re-create something similar.
Example:
test.shape
(200,)
test[0].shape
(5, 10)
test[1].shape
(3, 10)
test[2].shape
(6, 10)
The most reliable way of creating an object array is to initialize it and fill it. The behavior of np.array is too variable.
In [658]: alist = [np.ones((5,10),int), np.zeros((3,10),int), np.arange(60).resh
...: ape(6,10)]
In [659]: arr = np.empty(len(alist), dtype=object)
In [660]: arr[:] = alist
In [661]: arr
Out[661]:
array([ array([[1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1]]),
array([[0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]),
array([[ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9],
[10, 11, 12, 13, 14, 15, 16, 17, 18, 19],
[20, 21, 22, 23, 24, 25, 26, 27, 28, 29],
[30, 31, 32, 33, 34, 35, 36, 37, 38, 39],
[40, 41, 42, 43, 44, 45, 46, 47, 48, 49],
[50, 51, 52, 53, 54, 55, 56, 57, 58, 59]])], dtype=object)
The behavior of np.array varies with the relative shape of the components:
Nice object array:
In [668]: np.array((np.ones((3,5),int), np.ones((2,5),int)),object)
Out[668]:
array([array([[1, 1, 1, 1, 1],
[1, 1, 1, 1, 1],
[1, 1, 1, 1, 1]]),
array([[1, 1, 1, 1, 1],
[1, 1, 1, 1, 1]])], dtype=object)
3d array:
In [669]: np.array((np.ones((3,5),int), np.ones((3,5),int)),object)
Out[669]:
array([[[1, 1, 1, 1, 1],
[1, 1, 1, 1, 1],
[1, 1, 1, 1, 1]],
[[1, 1, 1, 1, 1],
[1, 1, 1, 1, 1],
[1, 1, 1, 1, 1]]], dtype=object)
error
In [670]: np.array((np.ones((3,4),int), np.ones((3,5),int)),object)
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-670-e9b41001d868> in <module>()
----> 1 np.array((np.ones((3,4),int), np.ones((3,5),int)),object)
ValueError: could not broadcast input array from shape (3,4) into shape (3)
You need to create an array with dtype=object, so that numpy knows to treat each entry as a Python object rather than treating the entire nested list as a single array.
For example:
import numpy as np
x = np.empty((5, 10))
y = np.empty((3, 10))
z = np.empty((6, 10))
test = np.array([x, y, z], dtype=object)
test[0].shape
# [5, 10]
x,y and z all have same shape? Or it's the second dimensions that differ?