[][src]Crate number_prefix

This is a library for formatting numbers with numeric prefixes, such as turning “3000 metres” into “3 kilometres”, or “8705 bytes” into “8.5 KiB”.

Usage

The function NumberPrefix::decimal returns either a pair of the resulting number and its prefix, or a notice that the number was too small to have any prefix applied to it. For example:

use number_prefix::{NumberPrefix, Standalone, Prefixed};

match NumberPrefix::decimal(8542_f32) {
    Standalone(bytes)   => println!("The file is {} bytes in size", bytes),
    Prefixed(prefix, n) => println!("The file is {:.0} {}B in size", n, prefix),
}

This will print out "The file is 8.5 kB in size". The {:.0} part of the formatting string tells it to restrict the output to only one decimal place. This value is calculated by repeatedly dividing the number by 1000 until it becomes less than that, which in this case results in 8.542, which gets rounded down. Because only one division had to take place, the function also returns the decimal prefix Kilo, which gets converted to its internationally-recognised symbol when formatted as a string.

If the value is too small to have any prefixes applied to it — in this case, if it’s under 1000 — then the standalone value will be returned:

use number_prefix::{NumberPrefix, Standalone, Prefixed};

match NumberPrefix::decimal(705_f32) {
    Standalone(bytes)   => println!("The file is {} bytes in size", bytes),
    Prefixed(prefix, n) => println!("The file is {:.0} {}B in size", n, prefix),
}

This will print out "The file is 705 bytes in size", having chosen the other path to follow. In this particular example, the user expects different formatting for both bytes and kilobytes: while prefixed values are given more precision, there’s no point using anything other than whole numbers for just byte amounts. This is why the function pays attention to values without any prefixes — they often need to be special-cased.

Binary Prefixes

This library also allows you to use the binary prefixes, which use the number 1024 (210) as the multiplier, rather than the more common 1000 (103). This uses the NumberPrefix::binary function. For example:

use number_prefix::{NumberPrefix, Standalone, Prefixed};

match NumberPrefix::binary(8542_f32) {
    Standalone(bytes)   => println!("The file is {} bytes in size", bytes),
    Prefixed(prefix, n) => println!("The file is {:.0} {}B in size", n, prefix),
}

This will print out "The file is 8.3 KiB in size". A kibibyte is slightly larger than a kilobyte, so the number is smaller in the result; but other than that, it works in exactly the same way, with the binary prefix being converted to a symbol automatically.

Which type of prefix should I use?

There is no correct answer this question! Common practice is to use the binary prefixes for numbers of bytes, while still using the decimal prefixes for everything else. Computers work with powers of two, rather than powers of ten, and by using the binary prefixes, you get a more accurate representation of the amount of data.

Prefix Names

If you need to describe your unit in actual words, rather than just with the symbol, import the PrefixNames trait, which adds methods to output the prefix in a variety of formats. For example:

use number_prefix::{NumberPrefix, Standalone, Prefixed, PrefixNames};

match NumberPrefix::decimal(8542_f32) {
    Standalone(bytes)   => println!("The file is {} bytes in size", bytes),
    Prefixed(prefix, n) => println!("The file is {:.0} {}bytes in size", n, prefix.lower()),
}

Re-exports

pub use Prefix::Kilo;
pub use Prefix::Mega;
pub use Prefix::Giga;
pub use Prefix::Tera;
pub use Prefix::Peta;
pub use Prefix::Exa;
pub use Prefix::Zetta;
pub use Prefix::Yotta;
pub use Prefix::Kibi;
pub use Prefix::Mibi;
pub use Prefix::Gibi;
pub use Prefix::Tebi;
pub use Prefix::Pebi;
pub use Prefix::Exbi;
pub use Prefix::Zebi;
pub use Prefix::Yobi;
pub use NumberPrefix::Standalone;
pub use NumberPrefix::Prefixed;

Enums

NumberPrefix

The result of trying to apply a prefix to a floating-point value.

Prefix

A numeric prefix, either binary or decimal.

Traits

Amounts

Traits for floating-point values for both the possible multipliers. They need to be Copy, have defined 1000 and 1024s, and implement a bunch of operators.

PrefixNames

Formatting methods for prefix, for when you want to output things other than just the short-hand symbols.