np.vectorize() for custom functions in NumPy - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of vectorized custom function with conditional logic

What is the output of this code that uses np.vectorize() on a custom function with a condition?

NumPy

import numpy as np

def classify(x):
    if x > 0:
        return 'pos'
    elif x < 0:
        return 'neg'
    else:
        return 'zero'

vec_classify = np.vectorize(classify)
arr = np.array([-2, 0, 3])
result = vec_classify(arr)
print(result)

A ['zero' 'neg' 'pos']

B ['neg' 'pos' 'zero']

C ['pos' 'zero' 'neg']

D ['neg' 'zero' 'pos']

💡 Hint

Think about how the function classifies negative, zero, and positive numbers.

The function returns 'neg' for negative numbers, 'zero' for zero, and 'pos' for positive numbers. The input array is [-2, 0, 3], so the output matches ['neg', 'zero', 'pos'].

❓ data_output

intermediate

Resulting array shape after vectorizing a function returning tuples

What is the shape of the output array when applying np.vectorize() to a function that returns a tuple for each input element?

NumPy

import numpy as np

def to_pair(x):
    return (x, x**2)

vec_to_pair = np.vectorize(to_pair)
arr = np.array([1, 2, 3])
result = vec_to_pair(arr)
print(result.shape)

A (2, 3)

B (3, 2)

C (3,)

D (2,)

💡 Hint

Check what type of array is returned when the function returns tuples.

The vectorized function returns an array of tuples, so the shape is (3,), a 1D array with 3 elements each being a tuple.

🔧 Debug

advanced

Identify the error in vectorizing a function with multiple arguments

What error does this code raise when using np.vectorize() on a function with two arguments but passing only one array?

NumPy

import numpy as np

def add(x, y):
    return x + y

vec_add = np.vectorize(add)
arr = np.array([1, 2, 3])
result = vec_add(arr)
print(result)

A TypeError: add() missing 1 required positional argument: 'y'

B No error, outputs [2 4 6]

C SyntaxError: invalid syntax

D ValueError: operands could not be broadcast together

💡 Hint

Check how many arguments the function expects versus how many are passed.

The function add expects two arguments, but vec_add is called with only one array, causing a TypeError about the missing argument.

🚀 Application

advanced

Using np.vectorize to apply a function that returns different types

What is the output of this code where the vectorized function returns either int or str based on input?

NumPy

import numpy as np

def process(x):
    if x % 2 == 0:
        return x // 2
    else:
        return 'odd'

vec_process = np.vectorize(process)
arr = np.array([1, 2, 3, 4])
result = vec_process(arr)
print(result)

A ['odd' '1' 'odd' '2']

B ['odd' 1 'odd' 2]

C ['odd' '1' 'odd' 2]

D ['odd' 1 'odd' '2']

💡 Hint

Consider how numpy arrays handle mixed types in vectorized functions.

The function returns mixed types (int and str), so np.vectorize creates an object dtype array. When printed, the output is ['odd' 1 'odd' 2], with integers shown without quotes.

🧠 Conceptual

expert

Why use np.vectorize instead of a for loop for custom functions?

Which statement best explains the main advantage of using np.vectorize() over a Python for loop when applying a custom function to a numpy array?

A np.vectorize allows applying functions element-wise with numpy syntax but does not guarantee speed improvement.

B np.vectorize automatically converts the function to use GPU acceleration.

C np.vectorize provides true parallel execution and speeds up computation significantly.

D np.vectorize compiles the function to machine code for faster execution.

💡 Hint

Think about what np.vectorize actually does under the hood.

np.vectorize is a convenience wrapper that applies a function element-wise using numpy syntax but does not provide true vectorized speed or parallelism.