NumPy ndarray.reshape() Method

NumPy ndarray.reshape()

The numpy.ndarray.reshape() method allows changing the shape of a NumPy array without altering its data. It returns a new view of the original array whenever possible, or a copy if required.

Syntax

ndarray.reshape(shape, /, *, order='C', copy=None)

Parameters

Parameter	Type	Description
shape	int or tuple of ints	The new shape for the array. It must be compatible with the original size.
order	{‘C’, ‘F’, ‘A’}	Memory layout order: – ‘C’ (row-major, C-style, default) – ‘F’ (column-major, Fortran-style) – ‘A’ (Fortran-like if input is Fortran-contiguous, otherwise C-like)
copy	bool, optional	Whether to create a copy of the data. Defaults to None, meaning it only copies if needed.

Return Value

Returns an array with the specified shape. If possible, it provides a view of the original array; otherwise, it creates a new copy.

---

Examples

1. Basic Reshaping of a 1D Array

Converting a 1D array into a 2D array of shape (3, 2).

import numpy as np

# Creating a 1D array with 6 elements
arr = np.array([1, 2, 3, 4, 5, 6])

# Reshaping into a 2D array with 3 rows and 2 columns
reshaped_arr = arr.reshape((3, 2))

print(reshaped_arr)

Output:

[[1 2]
 [3 4]
 [5 6]]

2. Using -1 to Auto-Calculate One Dimension

Using -1 lets NumPy automatically compute one dimension while maintaining the total size.

import numpy as np

# Original 1D array with 12 elements
arr = np.arange(12)

# Reshaping into a 3-row array with auto-computed columns
reshaped_arr = arr.reshape((3, -1))

print(reshaped_arr)

Output:

[[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]]

3. Reshaping with Fortran (‘F’) Order

Reshaping while filling elements column-wise using Fortran-style ordering.

import numpy as np

# Creating a 1D array with 6 elements
arr = np.array([1, 2, 3, 4, 5, 6])

# Reshaping with Fortran order (column-major)
reshaped_arr = arr.reshape((3, 2), order='F')

print(reshaped_arr)

Output:

[[1 4]
 [2 5]
 [3 6]]

Elements are filled column-wise instead of row-wise as in C-order.

4. Ensuring a Copy of the Array

Using the copy parameter to force a copy instead of a view.

import numpy as np

# Creating a 1D array
arr = np.array([1, 2, 3, 4, 5, 6])

# Reshaping with a forced copy
reshaped_arr = arr.reshape((2, 3), copy=True)

# Modifying the original array
arr[0] = 99

print("Original Array:")
print(arr)

print("Reshaped Array:")
print(reshaped_arr)

Output:

Original Array:
[99  2  3  4  5  6]

Reshaped Array:
[[1 2 3]
 [4 5 6]]

The reshaped array remains unchanged when modifying the original array, confirming that a copy was created.