np.vectorize() for custom functions in NumPy - Time & Space Complexity

Time Complexity: np.vectorize() for custom functions

O(n)

Understanding Time Complexity

We want to understand how the time it takes to run a custom function on many numbers changes when using np.vectorize().

How does the work grow when we apply the function to bigger arrays?

Scenario Under Consideration

Analyze the time complexity of the following code snippet.

import numpy as np

def my_func(x):
    return x ** 2 + 1

n = 10
vec_func = np.vectorize(my_func)
arr = np.arange(n)
result = vec_func(arr)

This code applies a custom function to each element of an array using np.vectorize.

Identify Repeating Operations

Identify the loops, recursion, array traversals that repeat.

• Primary operation: Applying my_func to each element of the array.
• How many times: Once for each element in the input array (n times).

How Execution Grows With Input

As the array size grows, the function runs once per element, so the total work grows in direct proportion to the number of elements.

Input Size (n)	Approx. Operations
10	10 function calls
100	100 function calls
1000	1000 function calls

Pattern observation: Doubling the input size roughly doubles the work done.

Final Time Complexity

Time Complexity: O(n)

This means the time to run grows linearly with the number of elements in the array.

Common Mistake

[X] Wrong: "np.vectorize makes the function run faster by doing things all at once like built-in numpy operations."

[OK] Correct: np.vectorize is a convenience tool that still calls the function once per element in Python, so it does not speed up the function itself.

Interview Connect

Understanding how np.vectorize works helps you explain performance when using custom functions on arrays, a useful skill for data tasks and coding interviews.

Self-Check

What if we replaced np.vectorize with a true numpy ufunc? How would the time complexity change?

Practice

1. What is the main purpose of np.vectorize() in numpy?

easy

A. To apply a custom function element-wise on numpy arrays without explicit loops

B. To create random numbers in numpy arrays

C. To sort numpy arrays quickly

D. To reshape numpy arrays into different dimensions

Solution

1. Step 1: Understand the role of np.vectorize()

   np.vectorize() takes a function that works on single values and makes it work on arrays element-wise.

2. Step 2: Compare with other numpy functions

   It is not used for random number generation, sorting, or reshaping arrays, which are handled by other numpy functions.

3. Final Answer:

   To apply a custom function element-wise on numpy arrays without explicit loops -> Option A

4. Quick Check:

   np.vectorize() = element-wise function application [OK]

Hint: Remember: vectorize applies functions element-wise on arrays [OK]

Common Mistakes:

• Confusing vectorize with random number generation
• Thinking vectorize reshapes arrays
• Assuming vectorize sorts arrays

2. Which of the following is the correct way to create a vectorized version of a function f using numpy?

easy

A. vfunc = np.vectorize[f]

B. vfunc = np.vectorize(f)

C. vfunc = np.vectorize.f()

D. vfunc = np.vectorize->f

Solution

1. Step 1: Recall the syntax of np.vectorize

   The correct syntax is to call np.vectorize as a function with the target function as argument: np.vectorize(f).

2. Step 2: Check the other options for syntax errors

   Options A, C, and D use incorrect syntax like square brackets, dot notation, or arrows which are invalid in Python.

3. Final Answer:

   vfunc = np.vectorize(f) -> Option B

4. Quick Check:

   Correct syntax uses parentheses: np.vectorize(f) [OK]

Hint: Use parentheses, not brackets or dots, with np.vectorize [OK]

Common Mistakes:

• Using square brackets instead of parentheses
• Trying to access vectorize as an attribute
• Using invalid arrow syntax

3. What will be the output of the following code?

import numpy as np
def square_plus_one(x):
    return x**2 + 1
vfunc = np.vectorize(square_plus_one)
arr = np.array([1, 2, 3])
result = vfunc(arr)
print(result)

medium

A. [2 5 10]

B. [1 4 9]

C. [2 3 4]

D. [1 2 3]

Solution

1. Step 1: Understand the function logic

   The function square_plus_one returns x**2 + 1. For inputs 1, 2, 3, the outputs are 1^2+1=2, 2^2+1=5, 3^2+1=10.

2. Step 2: Apply vectorized function on array

   Using np.vectorize, the function applies element-wise on the array [1, 2, 3], producing [2, 5, 10].

3. Final Answer:

   [2 5 10] -> Option A

4. Quick Check:

   Element-wise square plus one = [2 5 10] [OK]

Hint: Apply function to each element, then collect results [OK]

Common Mistakes:

• Forgetting to add 1 after squaring
• Confusing output with input array
• Assuming vectorize changes function logic

4. Identify the error in this code snippet:

import numpy as np
def add_five(x):
    return x + 5
vfunc = np.vectorize(add_five)
arr = [1, 2, 3]
result = vfunc(arr)
print result

medium

A. np.vectorize cannot be used with functions that add numbers

B. The function add_five should return x * 5 instead of x + 5

C. The print statement is missing parentheses

D. The input array should be a numpy array, not a list

Solution

1. Step 1: Check the print statement syntax

   In Python 3, print is a function and requires parentheses like print(result). print result causes a SyntaxError.

2. Step 2: Verify other parts of the code

   The function logic is correct, np.vectorize works fine with addition functions, and input lists are accepted by np.vectorize.

3. Final Answer:

   The print statement is missing parentheses -> Option C

4. Quick Check:

   print requires parentheses in Python 3 [OK]

Hint: print requires parentheses in Python 3 [OK]

Common Mistakes:

• Forgetting parentheses on print statement
• Assuming list input is invalid for vectorize
• Changing function logic unnecessarily

5. You want to apply a custom function that returns 'even' if a number is even and 'odd' if it is odd, on a numpy array. Which code correctly uses np.vectorize() to achieve this?

def even_odd(x):
    if x % 2 == 0:
        return 'even'
    else:
        return 'odd'
arr = np.array([1, 2, 3, 4])
# What next?

hard

A. result = np.apply_along_axis(even_odd, 1, arr)
print(result)

B. result = even_odd(arr)
print(result)

C. vfunc = np.vectorize(even_odd)
result = even_odd(arr.tolist())
print(result)

D. vfunc = np.vectorize(even_odd)
result = vfunc(arr)
print(result)

Solution

1. Step 1: Vectorize the custom function

   Use np.vectorize(even_odd) to create a vectorized version that works element-wise on arrays.

2. Step 2: Apply vectorized function on numpy array

   Call the vectorized function with the numpy array arr to get the result array of 'even'/'odd' strings.

3. Step 3: Check other options

   result = even_odd(arr) print(result) fails because even_odd does not support array input directly. vfunc = np.vectorize(even_odd) result = even_odd(arr.tolist()) print(result) fails because even_odd does not support list input. result = np.apply_along_axis(even_odd, 1, arr) print(result) fails with axis out of bounds for 1D array.

4. Final Answer:

   vfunc = np.vectorize(even_odd) result = vfunc(arr) print(result) -> Option D

5. Quick Check:

   Vectorize then apply on numpy array = vfunc = np.vectorize(even_odd) result = vfunc(arr) print(result) [OK]

Hint: Vectorize first, then apply on numpy array directly [OK]

Common Mistakes:

• Calling function directly on array without vectorize
• Calling non-vectorized function on list input
• Using invalid axis in apply_along_axis