6️⃣ NumPy ufunc (Universal Functions) – Perform Fast Vectorized Computation

Introduction – Why Learn NumPy ufunc?

NumPy’s ufuncs (universal functions) are fast, element-wise operations built in C and optimized for performance. Whether you’re working on arrays, matrices, or scientific datasets, ufuncs allow seamless arithmetic, logical, trigonometric, and set operations without writing slow loops.

In this guide, you’ll learn:

• What ufuncs are and how to use them efficiently
• How to create custom ufuncs with frompyfunc
• How to apply vectorized math like sum, diff, log, trig, and rounding
• Use cases of ufuncs in data processing and scientific computing

Topics Covered

NumPy ufunc Intro

Universal functions (ufunc) operate element-wise on arrays, enabling faster, more efficient code compared to loops.

import numpy as np

arr = np.array([1, 2, 3])
print(np.add(arr, 5))  # Output: [6 7 8]

Create Custom ufunc

Use np.frompyfunc() to turn any Python function into a ufunc.

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

ufunc_add = np.frompyfunc(custom_add, 2, 1)
print(ufunc_add([1, 2], [3, 4]))  # Output: [4 6]

Simple Arithmetic

a = np.array([1, 2, 3])
b = np.array([4, 5, 6])

print(np.add(a, b))
print(np.subtract(a, b))
print(np.multiply(a, b))
print(np.divide(b, a))

Rounding Decimals

x = np.array([1.7, 2.5, -3.2])

print(np.around(x))  # Nearest integer
print(np.floor(x))   # Round down
print(np.ceil(x))    # Round up

Logarithmic Functions

x = np.array([1, 2, 10, 100])

print(np.log(x))     # Natural log
print(np.log2(x))    # Base 2 log
print(np.log10(x))   # Base 10 log

Summations

x = np.array([1, 2, 3, 4])

print(np.sum(x))       # 10
print(np.cumsum(x))    # [1 3 6 10]

✖️ Products

x = np.array([1, 2, 3, 4])

print(np.prod(x))       # 24
print(np.cumprod(x))    # [1 2 6 24]

Differences

x = np.array([10, 15, 25, 35])

print(np.diff(x))       # [5 10 10]

Finding LCM

print(np.lcm.reduce([12, 18]))  # 36

Finding GCD

print(np.gcd.reduce([12, 18]))  # 6

Trigonometric Functions

x = np.array([0, 30, 60, 90])
radians = np.deg2rad(x)

print(np.sin(radians))
print(np.cos(radians))
print(np.tan(radians))

Hyperbolic Functions

x = np.array([0, 1, 2])

print(np.sinh(x))
print(np.cosh(x))
print(np.tanh(x))

Set Operations

a = np.array([1, 2, 3, 4])
b = np.array([3, 4, 5, 6])

print(np.union1d(a, b))         # [1 2 3 4 5 6]
print(np.intersect1d(a, b))     # [3 4]
print(np.setdiff1d(a, b))       # [1 2]
print(np.setxor1d(a, b))        # [1 2 5 6]

Summary – Recap & Next Steps

NumPy ufuncs are essential tools for high-performance numerical operations. With built-in support for element-wise math, logical processing, and custom functions, they replace slow loops and speed up workflows significantly.

Key Takeaways:

• ufuncs operate element-wise on NumPy arrays with C-speed performance.
• Functions like add, log, diff, sin, and gcd are built-in and optimized.
• Set operations and trigonometric/hyperbolic tools broaden their usage in science and engineering.

Real-World Relevance:
From scientific simulations to ML preprocessing and sensor data analysis, NumPy ufuncs help streamline operations in a fraction of the time.

FAQ – NumPy ufunc Basics

What is a ufunc in NumPy?

A ufunc is a fast element-wise function written in C that applies operations to arrays.

How is ufunc different from normal Python functions?

Ufuncs avoid explicit loops and use vectorized operations, making them significantly faster.

Can I create custom ufuncs?

Yes, using np.frompyfunc(func, nin, nout).

What is the use of np.diff()?

It calculates the difference between consecutive elements in an array.

Are ufuncs usable on multi-dimensional arrays?

Yes, all ufuncs support broadcasting and N-dimensional arrays.