CODEWITHSUNDEEP
pythonnumpy
user9673470
user9673470

Reputation: 79

I want to reshape 2D array into 3D array

I want to reshape 2D array into 3D array.I wrote codes,

for i in range(len(array)):
    i = np.reshape(i,(2,2,2))
    print(i)

i variable has even number's length array like [["100","150","2","4"],["140","120","3","5"]] or 

[[“1”,”5”,”6”,”2”],[“4”,”2”,”3”,”7”],[“7”,”5”,”6”,”6”],[“9”,”1”,”8”,”3”],[“3”,”4”,”5”,”6”],[“7”,”8”,”9”,”2”],,[“1”,”5”,”2”,”8”],[“6”,”7”,”2”,”1”],[“9”,”3”,”1”,”2”],[“6”,”8”,”3”,”3”]]

The length is >= 6. When I run this codes,ValueError: cannot reshape array of size 148 into shape (2,2,2) error happens. My ideal output is

[[['100', '150'], ['2', '4']], [['140', '120'], ['3', '5']]] or [[[“1”,”5”],[”6”,”2”]],[[“4”,”2”],[”3”,”7”]],[[“7”,”5”],[”6”,”6”]],[[“9”,”1”],[”8”,”3”]],[[“3”,”4”],[”5”,”6”]],[[“7”,”8”],[”9”,”2”]],[[“1”,”5”],[”2”,”8”]],[[“6”,”7”],[”2”,”1”]],[[“9”,”3”],[[”1”,”2”]],[[“6”,”8”],[”3”,”3”]]]

I rewrote the codesy = [[x[:2], x[2:]] for x in i] but output is not my ideal one.What is wrong in my codes?

Upvotes: 1

Views: 10447

Answers (3)

hpaulj
hpaulj

Reputation: 231738

In [76]: arr = np.arange(24).reshape(3,8)
In [77]: for i in range(len(arr)):
    ...:     print(i)
    ...:     i = np.reshape(i, (2,2,2))
    ...:     print(i)
    ...:     
0
....

AttributeError: 'int' object has no attribute 'reshape'

len(arr) is 3, so range(3) produces values, 0,1,2. You can't reshape the number 0.

Or did you mean to reshape arr[0], arr[1], etc?

In [79]: for i in arr:
    ...:     print(i)
    ...:     i = np.reshape(i, (2,2,2))
    ...:     print(i)
    ...:     
[0 1 2 3 4 5 6 7]
[[[0 1]
  [2 3]]

 [[4 5]
  [6 7]]]
[ 8  9 10 11 12 13 14 15]
[[[ 8  9]
  [10 11]]

 [[12 13]
  [14 15]]]
[16 17 18 19 20 21 22 23]
[[[16 17]
  [18 19]]

 [[20 21]
  [22 23]]]

That works - sort of. The prints look ok, but arr itself does not get changed:

In [80]: arr
Out[80]: 
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]])

That's because i is the iteration variable. Assigning a new value to it does not change the original object. If that's confusing, you need to review basic Python iteration.

Or we could iterate the range, and use it as an index:

In [81]: for i in range(len(arr)):
    ...:     print(i)
    ...:     x = np.reshape(arr[i], (2,2,2))
    ...:     print(x)
    ...:     arr[i] = x
    ...: 
    ...: 
    ...: 
    ...:     
0
[[[0 1]
  [2 3]]

 [[4 5]
  [6 7]]]
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-81-5f0985cb2277> in <module>()
      3     x = np.reshape(arr[i], (2,2,2))
      4     print(x)
----> 5     arr[i] = x
      6 
      7 

ValueError: could not broadcast input array from shape (2,2,2) into shape (8)

The reshape works, but you can't put a (2,2,2) array back into a slot of shape (8,). The number of elements is right, but the shape isn't.

In other words, you can't reshape an array piecemeal. You have to reshape the whole thing. (If arr was a list of lists, this kind of piecemeal reshaping would work.)

In [82]: np.reshape(arr, (3,2,2,2))
Out[82]: 
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]]]])

Upvotes: 0

user3483203
user3483203

Reputation: 51185

You don't need to loop to reshape the way you want to, just use arr.reshape((-1,2,2))

In [3]: x = np.random.randint(low=0, high=10, size=(2,4))

In [4]: x
Out[4]:
array([[1, 1, 2, 5],
       [8, 8, 0, 5]])

In [5]: x.reshape((-1,2,2))
Out[5]:
array([[[1, 1],
        [2, 5]],

       [[8, 8],
        [0, 5]]])

This approach will work for both of your arrays. The -1 as the first argument means numpy will infer the value of the unknown dimension.

Upvotes: 0

Trung Tran
Trung Tran

Reputation: 101

First of all, you are missing the meaning of reshaping. Let say your origin array has shape (A, B) and you want to reshape it to shape (M, N, O), you have to make sure that A * B = M * N * O. Obviously 148 != 2 * 2 * 2, right?

In your case, you want to reshape an array of shape (N, 4) to an array of shape (N, 2, 2). You can do like below:

x = np.reshape(y, (-1, 2, 2))

Hope this help :)

Upvotes: 3

Related Questions