How to round numbers in Python (round, floor, ceil)

How do you round a number in Python?

Short answer: use round(). round(3.78) returns 4, and round(3.14159, 2) returns 3.14. To always round up, use math.ceil(), and to always round down, use math.floor().

import math
 
print(round(3.78))         # 4
print(round(3.14159, 2))   # 3.14
print(math.ceil(3.21))     # 4  always up
print(math.floor(3.78))    # 3  always down

Watch out for exact halves: round(2.5) returns 2, not 3, because Python rounds halves to the nearest even number.

round() goes to the nearest value, halves go to the even number, math.ceil() rounds up and math.floor() rounds down.

Every way to round in Python

Round to the nearest integer or to decimals with round()

round(number, ndigits) takes two arguments:

• number: the value you want to round.
• ndigits (optional): how many decimals to keep. If you leave it out, you get an integer.

price = 19.4567
 
print(round(price))      # 19      int
print(round(price, 2))   # 19.46   float
print(round(price, 0))   # 19.0    float
print(round(1234.5678, -2))  # 1200.0

Without ndigits, round() returns an int. With ndigits, even 0, it returns a float. A negative ndigits rounds to tens, hundreds or thousands.

Why round(2.5) returns 2

Python rounds exact halves to the nearest even number. This is called banker's rounding, and it avoids pushing totals up when you round many values.

print(round(0.5))   # 0
print(round(1.5))   # 2
print(round(2.5))   # 2
print(round(3.5))   # 4
print(round(-2.5))  # -2

There is a second surprise with decimals: round(2.675, 2) returns 2.67, not 2.68. The number 2.675 cannot be stored exactly in binary, so Python actually has something slightly smaller, like 2.67499999....

If you need the rounding you learned at school, where halves always go up, use the decimal module with strings:

from decimal import Decimal, ROUND_HALF_UP
 
price = Decimal("2.675")
print(price.quantize(Decimal("0.01"), rounding=ROUND_HALF_UP))  # 2.68
print(Decimal("2.5").quantize(Decimal("1"), rounding=ROUND_HALF_UP))  # 3

Use Decimal for money, invoices and anything where every cent counts.

Round up and round down with math.ceil() and math.floor()

import math
 
print(math.ceil(3.21))    # 4
print(math.floor(3.78))   # 3
print(math.ceil(-3.2))    # -3
print(math.floor(-3.7))   # -4
print(math.trunc(-3.7))   # -3

To round up or down to decimals, multiply, round and divide:

print(math.ceil(3.14159 * 100) / 100)   # 3.15
print(math.floor(3.14159 * 100) / 100)  # 3.14

Round only when you show the number

Often you want to keep the full value and only show it rounded. Use an f-string:

total = 1234567.891
 
print(f"{total:.2f}")   # 1234567.89
print(f"{total:,.2f}")  # 1,234,567.89

The result is text, so do your calculations with the original number.

Round in NumPy and pandas

import numpy as np
import pandas as pd
 
print(np.round([1.234, 5.678], 1))           # [1.2 5.7]
print(pd.Series([1.234, 5.678]).round(1).tolist())  # [1.2, 5.7]

df.round(2) rounds every numeric column of a DataFrame. Both libraries also round halves to the nearest even number.

Use AI to check your rounding

An assistant like ChatGPT, Claude or GitHub Copilot can explain a surprising result and suggest the right method. These three prompts are ready to copy.

1. Explain a result

In Python, round([your number], [decimals]) returns [the result] and I expected [what you expected].
Explain why and show me how to get the result I want.

2. Choose a method

I need to round [prices / grades / measurements] in Python to [number] decimals,
always [up / down / to the nearest]. Show me the best method and explain why.

3. Round a dataset

I have a pandas DataFrame with these columns: [list the columns].
Show me how to round [the columns] to [number] decimals without changing the others.

Run every answer yourself and compare the results with a few values you know.

Common problems when rounding in Python

Related guides

• What is a float in Python
• How to multiply in Python
• Python double slash (//) operator
• How to comment out multiple lines in Python
• How to check your Python version