`import "github.com/abhinavdahiya/cockroach/util/decimal"`

- Constants
- func Cbrt(z, x *inf.Dec, s inf.Scale) *inf.Dec
- func Exp(z, n *inf.Dec, s inf.Scale) *inf.Dec
- func Float64FromDec(dec *inf.Dec) (float64, error)
- func Log(z *inf.Dec, x *inf.Dec, s inf.Scale) *inf.Dec
- func Log10(z *inf.Dec, x *inf.Dec, s inf.Scale) *inf.Dec
- func LogN(z *inf.Dec, x *inf.Dec, n *inf.Dec, s inf.Scale) *inf.Dec
- func Mod(z, x, y *inf.Dec) *inf.Dec
- func NewDecFromFloat(f float64) *inf.Dec
- func Pow(z, x, y *inf.Dec, s inf.Scale) *inf.Dec
- func PowerOfTenDec(pow int) *inf.Dec
- func PowerOfTenInt(pow int) *big.Int
- func SetFromFloat(z *inf.Dec, f float64) *inf.Dec
- func Sqrt(z, x *inf.Dec, s inf.Scale) *inf.Dec

const.go decimal.go loop.go table.go

❖

`const Precision = 16`

Precision defines the minimum precision all inexact decimal calculations should attempt to achieve.

Cbrt calculates the cube root of x to the specified scale and stores the result in z, which is also the return value.

The cube root calculation is implemented using Newton-Raphson method. We start with an initial estimate for cbrt(d), and then iterate:

x_{n+1} = 1/3 * ( 2 * x_n + (d / x_n / x_n) ).

Exp computes (e^n) (where n = a*b with a being an integer and b < 1) to the specified scale and stores the result in z, which is also the return value.

Float64FromDec converts a decimal to a float64 value, returning the value and any error that occurred. This converson exposes a possible loss of information.

Log computes the natural log of x using the Maclaurin series for log(1-x) to the specified scale and stores the result in z, which is also the return value. The function will panic if x is a negative number.

Log10 computes the log of x with base 10 to the specified scale and stores the result in z, which is also the return value. The function will panic if x is a negative number.

LogN computes the log of x with base n to the specified scale and stores the result in z, which is also the return value. The function will panic if x is a negative number or if n is a negative number.

Mod performs the modulo arithmatic x % y and stores the result in z, which is also the return value. It is valid for z to be nil, in which case it will be allocated internally. Mod will panic if the y is zero.

The modulo calculation is implemented using the algorithm:

x % y = x - (y * ⌊x / y⌋).

NewDecFromFloat allocates and returns a new Dec set to the given float64 value. The function will panic if the float is NaN or ±Inf.

Pow computes (x^y) as e^(y ln x) to the specified scale and stores the result in z, which is also the return value. If y is not an integer and x is negative nil is returned.

PowerOfTenDec returns an *inf.Dec with the value 10^pow. It should be treated as immutable.

Non-negative powers of 10 will have their value represented in their underlying big.Int with a scale of 0, while negative powers of 10 will have their value represented in their scale with a big.Int value of 1.

PowerOfTenInt returns a *big.Int with the value 10^pow. It should be treated as immutable.

SetFromFloat sets z to the given float64 value and returns z. The function will panic if the float is NaN or ±Inf.

Sqrt calculates the square root of x to the specified scale and stores the result in z, which is also the return value. The function will panic if x is a negative number.

The square root calculation is implemented using Newton's Method. We start with an initial estimate for sqrt(d), and then iterate:

x_{n+1} = 1/2 * ( x_n + (d / x_n) ).

Package decimal imports 6 packages (graph). Updated 2017-03-13. Refresh now. Tools for package owners. This is a dead-end fork (no commits since the fork).