np.newaxis for adding dimensions in NumPy - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of adding a new axis to a 1D array

What is the shape of the array b after adding a new axis using np.newaxis?

NumPy

import numpy as np

a = np.array([1, 2, 3])
b = a[:, np.newaxis]
print(b.shape)

A (3,)

B (3, 1)

C (1, 3)

D (1,)

💡 Hint

Think about how adding a new axis in the second position changes the shape.

Using np.newaxis in the second position converts the 1D array of shape (3,) into a 2D array with shape (3, 1).

❓ data_output

intermediate

Resulting array after adding new axis

What is the output of the array b after adding a new axis?

NumPy

import numpy as np

a = np.array([4, 5, 6])
b = a[np.newaxis, :]
print(b)

A [[4 5 6]]

B

[4
 5
 6]

C

[[4]
 [5]
 [6]]

D [4 5 6]

💡 Hint

Adding np.newaxis at the start adds a new row dimension.

Using np.newaxis at the start changes the shape from (3,) to (1, 3), making it a 2D array with one row.

❓ visualization

advanced

Visualizing shape changes with np.newaxis

Which option correctly describes the shape change when adding np.newaxis to a 2D array arr of shape (2, 3) as arr[:, :, np.newaxis]?

NumPy

import numpy as np

arr = np.array([[1, 2, 3], [4, 5, 6]])
new_arr = arr[:, :, np.newaxis]
print(new_arr.shape)

A (2, 3) - shape unchanged

B (2, 1, 3) - adds a new second dimension

C (1, 2, 3) - adds a new first dimension

D (2, 3, 1) - adds a new third dimension

💡 Hint

Adding np.newaxis at the end adds a new dimension at the last position.

The original array has shape (2, 3). Adding np.newaxis as the third index adds a new dimension at the end, resulting in shape (2, 3, 1).

🔧 Debug

advanced

Identify the error when using np.newaxis incorrectly

What error will this code raise?

NumPy

import numpy as np

a = np.array([1, 2, 3])
b = a[np.newaxis, np.newaxis]
print(b.shape)

A SyntaxError: invalid syntax

B TypeError: unhashable type: 'slice'

C (1, 1, 3) - no error, shape is correct

D IndexError: too many indices for array

💡 Hint

Check how many dimensions the original array has and how many indices are used.

The original array is 1D with shape (3,). Using two np.newaxis adds two new dimensions, resulting in shape (1, 1, 3). This is valid and does not raise an error.

🚀 Application

expert

Using np.newaxis to prepare data for broadcasting

Given two arrays a with shape (5,) and b with shape (3,), which option correctly uses np.newaxis to add dimensions so that a + b broadcasts to shape (5, 3)?

NumPy

import numpy as np

a = np.arange(5)
b = np.arange(3)

# Which option below correctly reshapes a and b for broadcasting?

A a[:, np.newaxis] + b[np.newaxis, :]

B a[np.newaxis, :] + b[:, np.newaxis]

C a[:, np.newaxis] + b[:, np.newaxis]

D a[np.newaxis, :] + b[np.newaxis, :]

💡 Hint

Broadcasting works when dimensions are compatible or equal to 1.

Adding np.newaxis to a as a[:, np.newaxis] changes shape from (5,) to (5, 1). Adding np.newaxis to b as b[np.newaxis, :] changes shape from (3,) to (1, 3). These shapes broadcast to (5, 3).