Numbers, Decimals, and Precision

Choose numeric types deliberately: avoid float for money, use Decimal with explicit rounding rules, and document units.

On this page

Numeric Types in Real Systems

int is exact. float is approximate. For money and billing, use Decimal with explicit rounding.

Float Surprise

print(0.1 + 0.2)  # 0.30000000000000004

Decimal for Money

from decimal import Decimal, ROUND_HALF_UP

price = Decimal("19.99")
tax_rate = Decimal("0.20")

total = (price * (Decimal("1") + tax_rate)).quantize(
    Decimal("0.01"),
    rounding=ROUND_HALF_UP,
)

Operational Checklist

No floats in money paths (pricing, invoices, payouts).
Rounding policy is defined and tested.
Units are explicit in code and configs (seconds vs milliseconds).

Failure Modes

Precision drift: small errors accumulate across computations.
Implicit casts: mixing int/float silently changes results.

← Strings and f-strings (Real-World Formatting)

Lists vs Tuples vs Sets (When to Use Which) →