This is one of the most straightforward errors in Python, often referred to as a ZeroDivisionError. It’s not a bug in the language; it’s a fundamental rule of mathematics. You cannot divide a number by zero.

# WRONG — denominator is zero, crashes on / and % equally
total_score = 100
num_students = 0
average = total_score / num_students    # ZeroDivisionError: division by zero

# FIX 1 — check the denominator before dividing (LBYL: Look Before You Leap)
if num_students > 0:
    average = total_score / num_students
else:
    average = 0    # or None, or float('nan') depending on your use case

# FIX 2 — catch the error at runtime (EAFP: Easier to Ask Forgiveness)
try:
    average = total_score / num_students
except ZeroDivisionError:
    average = 0

# FIX 3 — Pandas: use np.errstate to silence zero division in vectorised ops
import numpy as np
with np.errstate(divide='ignore', invalid='ignore'):
    result = np.where(denominator != 0, numerator / denominator, 0)

The Cause

Your code tried to perform a division (/) or modulo (%) operation where the denominator (the number on the bottom) was 0.

Problem Code:

total_score = 100
num_students = 0

# CRASH!
average = total_score / num_students

The Fix 1: The “Look Before You Leap” (LBYL) Check

The safest way is to check the denominator before you try to divide.

total_score = 100
num_students = 0

if num_students > 0:
    average = total_score / num_students
else:
    average = 0 # Or 'None', or 'N/A'

print(average)
# Output: 0 (No crash)

The Fix 2: The “Easier to Ask Forgiveness” (EAFP) try/except

This is the more “Pythonic” way. You just try the operation, and if it fails, you “catch” the ZeroDivisionError. This is perfect if you don’t expect it to happen often.

total_score = 100
num_students = 0

try:
    average = total_score / num_students
except ZeroDivisionError:
    average = 0

print(average)
# Output: 0 (No crash)

This is a much cleaner way to handle potential errors from user input or external data.

ZeroDivisionError: division by zero — LBYL vs EAFP and the Pandas Vectorised Fix

ZeroDivisionError: division by zero fires on /, //, and % when the right-hand operand is zero. Python stops immediately — no result, no infinity, no NaN. Just a crash.

Two fix philosophies exist. Pick based on how often you expect zero to appear.

Use the if check (LBYL) when zero is a predictable, normal case in your data — empty lists, fresh counters, survey responses with no participants. The guard runs on every call. Zero costs nothing in a simple script, but adds overhead in tight loops over large datasets.

Use try/except ZeroDivisionError (EAFP) when zero is rare and exceptional — a formula that should almost never divide by zero but might on bad input. The try block adds no overhead when zero doesn’t occur. It only costs time when the exception actually fires, making it faster than an if check on the happy path.

LBYL — best when zero is common

result = numerator / denominator if denominator != 0 else 0

EAFP — best when zero is rare

try:
result = numerator / denominator
except ZeroDivisionError:
result = 0

The Pandas and NumPy case needs neither. Division across a column or array fires RuntimeWarning and produces inf or NaN by default — it doesn’t raise ZeroDivisionError. Suppress the warning and control the output with np.where:

import numpy as np
result = np.where(df[‘count’] != 0, df[‘total’] / df[‘count’], 0)

This applies the division only where the denominator isn’t zero, substitutes 0 everywhere else, and runs at full NumPy vectorisation speed — no Python loop, no exception handling overhead.