ZeroDivisionError: division by zero — how to fix

ZeroDivisionError: division by zero
ZeroDivisionError: float division by zero
ZeroDivisionError: integer division or modulo by zero

Quick fix: Check the divisor before dividing: avg = total / count if count else 0. For averages of possibly-empty lists use statistics.fmean(xs) if xs else None. In NumPy/pandas division by zero gives inf/nan instead of raising, so handle those separately.

What it means

Python raises ZeroDivisionError whenever you divide or take a modulo by zero:

10 / 0      # ZeroDivisionError: division by zero
10.0 / 0.0  # ZeroDivisionError: float division by zero
10 // 0     # ZeroDivisionError: integer division or modulo by zero
10 % 0      # same

The real question is why the divisor is zero. The traceback shows the line; print the divisor just before it.

Cause 1: average or percentage of an empty collection

def average(xs):
    return sum(xs) / len(xs)   # crashes when xs == []

Fix 1: guard the empty case explicitly

def average(xs):
    return sum(xs) / len(xs) if xs else None

pct = done / total * 100 if total else 0.0

Decide what "no data" should mean (0, None, float("nan")) rather than hiding it. statistics.mean([]) raises StatisticsError instead, which is also explicit.

Cause 2: integer division truncated something to zero

step = width // count    # 5 // 10 == 0
x = 100 / step           # ZeroDivisionError

Use true division (/) when you need fractions, or enforce a minimum: step = max(1, width // count).

Cause 3: a value defaulted to 0

Config values, dict lookups with .get(key, 0), failed parsing, or a database column with 0 for "unknown":

rate = config.get("rate", 0)
period = 1 / rate

Validate inputs where they enter the program and raise a clear error:

if rate <= 0:
    raise ValueError(f"rate must be positive, got {rate!r}")

Cause 4: elapsed time is zero

Measuring speed with time.time() over a very fast operation can give 0 elapsed seconds on some platforms. Use time.perf_counter() (high resolution) and still guard: rate = n / elapsed if elapsed > 0 else float("inf").

Fix: catching it

When zero is an expected, rare case, try/except is fine:

try:
    ratio = a / b
except ZeroDivisionError:
    ratio = float("nan")

Avoid wrapping big blocks in except ZeroDivisionError — you'll hide bugs in unrelated lines.

NumPy and pandas behave differently

They don't raise; they return inf, -inf or nan and NumPy prints RuntimeWarning: divide by zero encountered:

import numpy as np
np.array([1, 2]) / np.array([0, 1])     # array([inf, 2.])

with np.errstate(divide="ignore", invalid="ignore"):
    r = np.divide(a, b, out=np.zeros_like(a, dtype=float), where=b != 0)

In pandas:

df["ratio"] = (df["a"] / df["b"]).replace([np.inf, -np.inf], np.nan)

A NumPy scalar like np.float64(1) / 0 also returns inf with a warning instead of raising, which can be surprising when values come from a DataFrame.

Decimal

decimal.Decimal(1) / 0 raises decimal.DivisionByZero (a subclass of ZeroDivisionError), so the same except works.

Last updated 2026-11-04.