TypeError: 'float' object cannot be interpreted as an integer — how to fix

Traceback (most recent call last):
  File "main.py", line 3, in <module>
    for i in range(n / 2):
TypeError: 'float' object cannot be interpreted as an integer

Quick fix: Use integer division // instead of / (range(n // 2)), or convert explicitly with int(x), round(x) or math.ceil(x). For NumPy values use int(arr.shape[0] / 2) or //.

What it means

Some built-ins only accept integers: range(), list * n, str * n, bytes(n), chr(), hex(), slice indices, random.randrange() and others. They call the object's __index__ method, which float deliberately doesn't have, because 2.5 has no exact integer meaning. Even 4.0 is rejected.

range(5.0)      # TypeError
"-" * 3.0       # TypeError: can't multiply sequence by non-int of type 'float'

Cause 1: / returns a float in Python 3

In Python 2, 7 / 2 was 3. In Python 3 it is 3.5, and 10 / 2 is 5.0 — still a float.

n = 10
for i in range(n / 2):     # range(5.0) -> TypeError
    ...

Fix 1: use floor division //

for i in range(n // 2):    # 5
    ...

mid = len(items) // 2
left, right = items[:mid], items[mid:]

// returns an int when both operands are ints. Note 7.0 // 2 is still 3.0 (a float), so if either side might be a float, convert explicitly.

Fix 2: convert explicitly, choosing the rounding you want

import math

int(3.7)        # 3   (truncates toward zero)
round(3.7)      # 4   (banker's rounding: round(2.5) == 2)
math.floor(3.7) # 3
math.ceil(3.2)  # 4   (e.g. number of pages)

pages = math.ceil(total_items / per_page)
for p in range(pages):
    ...

Pick deliberately: int() silently drops the fraction, which can hide off-by-one bugs.

Cause 2: values that came from parsing or math functions

float(input()), json.load() with 10.0 in the file, math.sqrt(), math.pow() and ** with a negative or float exponent all return floats:

side = math.sqrt(area)        # 4.0, a float
for row in range(side):       # TypeError

Fix: range(int(side)), or use math.isqrt(area) (Python 3.8+), which returns an int.

If the value comes from user input and should be a whole number, parse it as one: n = int(input()).

Cause 3: NumPy and pandas values

NumPy floats (numpy.float64) behave like float here:

import numpy as np
arr = np.zeros(10)
half = arr.size / 2            # 5.0
arr[:half]                     # TypeError: slice indices must be integers
np.zeros(half)                 # TypeError: 'float' object cannot be interpreted as an integer

Fix with arr.size // 2 or int(...). For np.linspace(0, 1, num), num must also be an int.

pandas columns with missing values are stored as float64 (NaN is a float), so df["count"].max() may return 12.0. Convert with int(...), or use the nullable integer dtype: df["count"] = df["count"].astype("Int64").

Cause 4: random and time helpers

random.randrange(0, 10.0) and random.sample(population, k=2.0) raise the same family of errors in recent Python versions. Pass ints.

Finding it quickly

The traceback's last line points at the call. Print the type of the suspicious argument:

print(type(n), n)

If it says <class 'float'>, trace it back to the first /, float(), math.* call or JSON value that produced it, and fix it there rather than sprinkling int() everywhere.

Last updated 2026-11-06.