Python ZeroDivisionError: Division by Zero

3ZeroDivisionError: explained step by step

A ZeroDivisionError means your code asked Python to divide by zero, or to take a remainder after dividing by zero. Mathematics has no answer for that, so Python stops instead of inventing one.

Which operations raise ZeroDivisionError

The message changes slightly with the operator, but the cause is always a zero on the right-hand side. All of Python's division-style operators are affected: /, //, % and divmod().

operations = {
    "10 / 0": lambda: 10 / 0,
    "10 // 0": lambda: 10 // 0,
    "10 % 0": lambda: 10 % 0,
    "10 / 0.0": lambda: 10 / 0.0,
    "divmod(10, 0)": lambda: divmod(10, 0),
    "0 ** -1": lambda: 0 ** -1,
}

for text, operation in operations.items():
    try:
        operation()
    except ZeroDivisionError as err:
        print(f"{text:14} -> {err}")

Notice that 10 / 0.0 fails too. Unlike some languages, Python's floats do not quietly produce infinity when you divide by zero. Libraries such as NumPy do return inf (with a warning), but plain Python always raises.

How to read the traceback

Traceback (most recent call last):
  File "stats.py", line 6, in <module>
    print(average([]))
  File "stats.py", line 3, in average
    return sum(numbers) / len(numbers)
           ~~~~~~~~~~~~~^~~~~~~~~~~~~~
ZeroDivisionError: division by zero

1. The last line shows the operator in use, division by zero.
2. The innermost frame (the last File entry) is where the division happens: sum(numbers) / len(numbers).
3. The frame above it shows the call that supplied the bad value: average([]). The bug is usually here, in what you passed in, not in the division itself.

Cause 1: averaging an empty list

The most common real-world cause: dividing by len(...) of something that turned out to be empty, such as a filter with no matches.

def average(numbers):
    return sum(numbers) / len(numbers)

print(average([4, 8, 6]))
try:
    print(average([]))
except ZeroDivisionError as err:
    print("ZeroDivisionError:", err)

Cause 2: the user (or a data file) supplies zero

Anything that comes from outside your program can be zero: a form field, a CSV column, an API value. Convert it, then decide what zero should mean. See ValueError for the related problem of text that is not a number at all.

Cause 3: a computed denominator that ends up zero

ratio = passed / (total - failed) looks harmless until every test fails. When the divisor is itself a calculation, think about which inputs can make it zero.

Fix 1: guard with an if statement

Check the divisor first when zero is an expected situation, and choose what the result should be. The if/else lesson covers the syntax.

def safe_average(numbers):
    if len(numbers) == 0:
        return None          # no data, so no average
    return sum(numbers) / len(numbers)

print(safe_average([70, 90]))
print(safe_average([]))

Return None, 0 or a message: which one is right depends on your program. An average of nothing is not zero, so None is often more honest.

Fix 2: catch the error with try/except

Use try and except when zero is rare, or when the check would be awkward to write. Catch ZeroDivisionError specifically, never a bare except, so that real bugs still show up.

def split_bill(total, people):
    try:
        return round(total / people, 2)
    except ZeroDivisionError:
        return "Add at least one person"

print(split_bill(90, 4))
print(split_bill(90, 0))

Fix 3: validate input before the maths

When the zero comes from a user, reject it early with a clear message, ideally before any calculation. This keeps the division line simple and puts the explanation where the user can act on it.

Zero is not the same as "not a number"

The error only concerns an exact zero divisor. Using float("inf") or math.inf gives you a value that behaves like infinity without raising anything, and float("nan") represents "not a number". Neither is a replacement for checking your inputs, because they hide the problem rather than solve it. For statistics, the math and statistics lesson shows statistics.mean([]) raising a clearer StatisticsError for empty data.

If your message instead says unsupported operand type(s) for /, that is a TypeError: the types are wrong, not the value.