Broadcasting compatibility check in NumPy - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of broadcasting shape check

What is the output of the following code that checks if two arrays can be broadcast together?

NumPy

import numpy as np
shape1 = (3, 1, 5)
shape2 = (1, 4, 5)
result = np.broadcast_shapes(shape1, shape2)
print(result)

A (3, 4, 5)

B (3, 1, 5, 4, 5)

C ValueError: shape mismatch: objects cannot be broadcast to a single shape

D (3, 4)

💡 Hint

Broadcasting compares shapes from the right side, matching dimensions or using 1 to expand.

The shapes (3,1,5) and (1,4,5) broadcast to (3,4,5) because 1 can expand to match the other dimension.

❓ Predict Output

intermediate

Broadcasting compatibility check result

What happens when you try to broadcast these shapes using numpy.broadcast_shapes?

NumPy

import numpy as np
shape1 = (2, 3)
shape2 = (3, 2)
try:
    result = np.broadcast_shapes(shape1, shape2)
except Exception as e:
    result = type(e).__name__
print(result)

A (2, 3, 3, 2)

B ValueError

C (3, 3)

D TypeError

💡 Hint

Broadcasting requires dimensions to be equal or one of them to be 1 from the right side.

Shapes (2,3) and (3,2) are incompatible for broadcasting because their trailing dimensions differ and neither is 1.

❓ data_output

advanced

Resulting shape after broadcasting arrays

Given two arrays with shapes (4, 1, 6) and (3, 6), what is the shape of the result after broadcasting?

NumPy

import numpy as np
arr1 = np.zeros((4, 1, 6))
arr2 = np.zeros((3, 6))
result = np.broadcast(arr1, arr2)
print(result.shape)

A (4, 1, 6, 3, 6)

B (3, 6)

C ValueError

D (4, 3, 6)

💡 Hint

Broadcasting aligns shapes from the right and expands dimensions of size 1.

The shape (3,6) is treated as (1,3,6) to align with (4,1,6), resulting in (4,3,6).

🔧 Debug

advanced

Identify the broadcasting error

Which option shows the code that will raise a broadcasting error when adding two numpy arrays?

A

import numpy as np
x = np.ones((2, 3))
y = np.ones((3, 2))
result = x + y

B

import numpy as np
x = np.ones((5, 1))
y = np.ones((1, 4))
result = x + y

C

import numpy as np
x = np.ones((3, 1, 2))
y = np.ones((1, 4, 2))
result = x + y

D

import numpy as np
x = np.ones((1, 3))
y = np.ones((3, 1))
result = x + y

💡 Hint

Check if dimensions align from the right or are 1 for broadcasting.

Shapes (2,3) and (3,2) cannot broadcast because their trailing dimensions differ and neither is 1.

🚀 Application

expert

Determine output shape for complex broadcasting

You have arrays with shapes (7, 1, 5, 1) and (1, 3, 1, 4). What is the shape of the result after broadcasting these arrays?

NumPy

import numpy as np
shape1 = (7, 1, 5, 1)
shape2 = (1, 3, 1, 4)
result_shape = np.broadcast_shapes(shape1, shape2)
print(result_shape)

A (7, 1, 5, 4)

B (1, 3, 5, 1)

C (7, 3, 5, 4)

D ValueError

💡 Hint

Broadcasting expands dimensions where one shape has 1 and the other has a size >1.

Each dimension is compared from right to left: 1 vs 4 → 4, 5 vs 1 → 5, 1 vs 3 → 3, 7 vs 1 → 7.