Reshape Array in NumPy

NumPy – Reshape Array

To reshape a given array to specific shape using NumPy library, we can use numpy.reshape() function.

Pass the given array, and required shape (as tuple) as arguments to the numpy.reshape() function.

Examples

1. Reshape numpy array from (3,4) to (2,6)

In the following program, we reshape a numpy array of shape (3, 4) to (2, 6).

import numpy as np

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

shape = (6, 2)
output = np.reshape(arr, shape)

print(output)

Output

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

2. Reshape numpy array from (3,4) to (-1,2)

If any of the dimension in the input shape is given as -1, then this dimension is adjusted based on the length in other dimensions.

For example, in the following program, we reshape a numpy array of shape (3, 4) to (-1, 2). Since we have -1 for the first dimension in the output shape, that dimension’s length is computed from (3*4)/(2) which is 6.

import numpy as np

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

shape = (-1, 2)
output = np.reshape(arr, shape)

print(output)

Output

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

Summary

In this NumPy Tutorial, we learned how to reshape a given numpy array in Python using numpy.reshape() function.