NumPy Array Reshape

Reshaping Arrays

Reshaping means changing the shape of an array - the number of elements along each dimension - without changing the underlying data. As long as the total element count matches, you can rearrange a flat array into rows and columns, or a 2-D array into a stack of smaller 2-D blocks.

Reshape From 1-D to 2-D

import numpy as np

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

newarr = arr.reshape(4, 3)
print(newarr)

# Output:
# [[ 1  2  3]
#  [ 4  5  6]
#  [ 7  8  9]
#  [10 11 12]]

Reshape From 1-D to 3-D

import numpy as np

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

newarr = arr.reshape(2, 3, 2)
print(newarr)

# Output:
# [[[ 1  2]
#   [ 3  4]
#   [ 5  6]]
#
#  [[ 7  8]
#   [ 9 10]
#   [11 12]]]

The new shape must account for exactly the same number of elements as the original array. Reshaping 12 elements into (4, 3) works because 4 x 3 = 12, but reshape(3, 3) raises a ValueError because 3 x 3 = 9 does not match 12.

Unknown Dimension With -1

You do not always have to specify every dimension yourself. Pass -1 for one dimension and NumPy calculates its correct size automatically based on the array's length and the dimensions you did specify. You can only use -1 for a single dimension at a time.

import numpy as np

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

newarr = arr.reshape(2, 2, -1)
print(newarr)
print(newarr.shape)

# Output:
# [[[ 1  2  3]
#   [ 4  5  6]]
#
#  [[ 7  8  9]
#   [10 11 12]]]
# (2, 2, 3)