np.reshape() returns a new view (or, when necessary, a copy) of an array with a different shape, without changing the underlying data or the total number of elements.

Syntax

np.reshape(arr, newshape)
arr.reshape(newshape)

Deep Dive Course

reshape only reorganizes how the same elements are grouped into dimensions — the total element count must stay exactly the same, so a 12-element array can become (3, 4), (4, 3), (2, 6), or (2, 2, 3), but never (3, 5). Passing -1 for one dimension tells NumPy to calculate that dimension automatically from the array's total size and the other specified dimensions, which is convenient when you know some dimensions but not all of them.

1Understanding np.reshape()

💡

Use -1 for one dimension in reshape(), e.g. arr.reshape(-1, 1) to make a column vector, instead of computing that dimension yourself — it's less error-prone and adapts automatically if the array's size changes.

editor.html

import numpy as np

arr = np.arange(12)
reshaped = arr.reshape(3, 4)
print(reshaped)

localhost:3000

2Practical Example

Here is a real-world application of np.reshape() showing how it is used in production NumPy code.

editor.html

import numpy as np

arr = np.arange(6)
column = arr.reshape(-1, 1)
print(column)

localhost:3000

3Best Practices

Follow these guidelines when working with np.reshape():

1. Use -1 for exactly one dimension in reshape() to let NumPy infer it, instead of computing and hardcoding that value yourself

2. Remember reshape() returns a view when possible, so mutating the reshaped array can also mutate the original — call .copy() if you need independence

3. Check the total element count matches before reshaping, or catch the resulting ValueError, rather than assuming a reshape will always succeed

⚠️

Tip: Use -1 for one dimension in reshape(), e.g. arr.reshape(-1, 1) to make a column vector, instead of computing that dimension yourself — it's less error-prone and adapts automatically if the array's size changes.

editor.html

import numpy as np

arr = np.arange(12)
reshaped = arr.reshape(3, 4)
print(reshaped)

localhost:3000

Examples

Example 01Basic Usage

import numpy as np

arr = np.arange(12)
reshaped = arr.reshape(3, 4)
print(reshaped)

Example 02Advanced Example

import numpy as np

arr = np.arange(6)
column = arr.reshape(-1, 1)
print(column)

Best Practices

Use -1 for exactly one dimension in reshape() to let NumPy infer it, instead of computing and hardcoding that value yourself
Remember reshape() returns a view when possible, so mutating the reshaped array can also mutate the original — call .copy() if you need independence
Check the total element count matches before reshaping, or catch the resulting ValueError, rather than assuming a reshape will always succeed

Interview Question

What error does NumPy raise if you try to reshape a 10-element array into a shape of (3, 4), and why?▼

Hint: Think about the constraint reshape() must preserve.

NumPy raises a ValueError, because reshape() requires the new shape to have exactly the same total number of elements as the original array — 3 times 4 is 12, which doesn't match the original 10 elements, so there's no way to redistribute the existing data into that shape without losing or fabricating values. reshape() only ever reorganizes existing elements; it never adds or removes any.

Exercises

MediumPractice using np.reshape() in a real scenario.

View Solution

import numpy as np

arr = np.arange(12)
reshaped = arr.reshape(3, 4)
print(reshaped)

Frequently Asked Questions

What error does NumPy raise if you try to reshape a 10-element array into a shape of (3, 4), and why?▼

Related Functions

ndarray-shapenp-ravelndarray-flatten