np.frompyfunc() for ufunc creation in NumPy - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of a simple ufunc created with np.frompyfunc()

What is the output of this code snippet that creates a ufunc to add 10 to each element?

NumPy

import numpy as np

def add_ten(x):
    return x + 10

ufunc_add_ten = np.frompyfunc(add_ten, 1, 1)
result = ufunc_add_ten(np.array([1, 2, 3]))
print(result)

A TypeError

B [1 2 3 10]

C [11 12 13]

D ['11' '12' '13']

💡 Hint

np.frompyfunc returns an object array by default.

np.frompyfunc creates a ufunc that returns an object array containing integer values, printed as ['11' '12' '13'].

❓ data_output

intermediate

Result shape and type from a ufunc with multiple outputs

Given this ufunc that returns two values per input, what is the shape and type of the output when applied to a 1D array of length 3?

NumPy

import numpy as np

def square_and_cube(x):
    return x**2, x**3

ufunc_sc = np.frompyfunc(square_and_cube, 1, 2)
result = ufunc_sc(np.array([2, 3, 4]))
print(type(result))
print(len(result))
print(result[0])

A Type: tuple, Length: 2, First element: [4 9 16]

B TypeError

C Type: tuple, Length: 2, First element: [4 9 16] as object array

D Type: numpy.ndarray, Length: 3, First element: 4

💡 Hint

np.frompyfunc returns object arrays even for multiple outputs.

The ufunc returns a tuple of two object arrays, each containing the squared and cubed values respectively.

🔧 Debug

advanced

Identify the error in ufunc creation with np.frompyfunc()

What error will this code raise and why?

NumPy

import numpy as np

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

ufunc_faulty = np.frompyfunc(faulty_func, 1, 1)

A TypeError: Number of input arguments does not match function signature

B No error, ufunc created successfully

C ValueError: Invalid number of inputs for frompyfunc

D SyntaxError

💡 Hint

Check the number of input arguments specified vs the function definition.

No error is raised because np.frompyfunc does not inspect or validate the Python function's signature against nin at creation time. The ufunc is created successfully, but a TypeError would occur upon calling it due to argument count mismatch.

🚀 Application

advanced

Using np.frompyfunc() to vectorize a string operation

You want to create a ufunc that converts strings to uppercase. Which option correctly creates and applies this ufunc to a numpy array of strings?

NumPy

import numpy as np

# Your code here

result = ufunc_upper(np.array(['apple', 'banana', 'cherry']))
print(result)

A ufunc_upper = np.frompyfunc(lambda s: s.upper(), 1, 1)

B ufunc_upper = np.frompyfunc(str.upper, 1, 1)

C ufunc_upper = np.frompyfunc(lambda s: s.upper, 1, 1)

D ufunc_upper = np.frompyfunc(lambda s: s.upper, 0, 1)

💡 Hint

Remember to call the method with parentheses to execute it.

Option A correctly defines a lambda that calls upper() on the string, matching nin=1 and nout=1.

🧠 Conceptual

expert

Understanding output types of np.frompyfunc() ufuncs

Which statement about the output of ufuncs created with np.frompyfunc() is TRUE?

A They return object arrays regardless of input types, because they wrap Python functions.

B They always return numpy arrays with the same dtype as the input arrays.

C They return lists instead of numpy arrays.

D They return arrays with dtype inferred from the function's return type.

💡 Hint

np.frompyfunc wraps Python functions and cannot infer fixed dtypes.

np.frompyfunc creates ufuncs that return object arrays because the underlying Python function can return any type.