Data Types: Decimal

fn main() {
    const price: Decimal = 499.99
    const quantity: Decimal = 3
    print(price * quantity)
    const huge: Decimal = 99999999999999999999999999999999999999.99
    const tiny: Decimal = 0.000000000000000000000000000001
    print(huge + tiny)
}

1499.9799999999999999999999999999999999999999.990000000000000000000000000001

Write the type, : Decimal, and the number is kept exactly as written. A Float would compute 0.1 + 0.2 as 0.30000000000000004 when worked out while the program runs; a Decimal gives 0.3:

fn main() {
    const a = 0.1 + 0.2
    print(a)
    let x = 0.1
    print(x + 0.2)
    const price: Decimal = 0.1
    const total = price + Decimal(0.2)
    print(total)
}

0.30.300000000000000040.3

Rounding

fn main() {
    const a: Decimal = 2.345
    const b: Decimal = -2.345
    const c: Decimal = 2.355
    print(a.round(2))
    print(b.round(2))
    print(a.round_even(2))
    print(c.round_even(2))
}

2.35-2.352.342.36

• round(2) rounds to 2 places, a half away from zero: 2.345 becomes 2.35, -2.345 becomes -2.35. This is how most people round by hand.
• round_even(2) rounds a half to the even digit, "banker's rounding": 2.345 becomes 2.34, 2.355 becomes 2.36. Over many sums, halves then go up as often as down, so totals don't drift upward. Banks and statistics use it.

As big as it needs to be

There is no separate BigDecimal type to learn: a Decimal grows by itself. Up to about 38 digits it is kept in a fast 128-bit form, and beyond that it keeps every digit, with no change to your code. A division that never ends keeps 18 digits after the point.