TinyFloat

TinyFloat is a tiny decimal class for JavaScript.

It solves the main problem with the IEEE 754 floats (the notorious 0.1 + 0.2 == 0.30000000000000004) without implementing an arbitrary-precision type that is more accurate but also heavier.

TinyFloat is just 0.5 kB and has no dependencies, while decimal.js is 12.3 kB. In most cases, including working with money, the precision provided by decimal.js is irrelevant, so TinyFloat is an excellent alternative.

Installation

The library is available as an npm package:

npm install tinyfloat --save

Usage

The library exports the TinyFloat class, which you can use to create decimal numbers:

import { TinyFloat } from "tinyfloat";

new TinyFloat("0.1").add("0.2").toNumber();
//=> 0.3

You can also set the precision for the TinyFloat instance:

new TinyFloat("0.123456", 2).toNumber();
//=> 0.12

new TinyFloat("0.987654321", 2).toNumber();
//=> 0.99

API

The TinyFloat class has basic arithmetic methods:

• add - adds two numbers
• sub - subtracts two numbers
• mul - multiplies two numbers
• div - divides two numbers
• mod - returns the remainder of the division

All methods accept strings, numbers, and TinyFloat instances:

new TinyFloat("0.1").add(new TinyFloat("0.2")).toNumber();
// Or pass a string:
new TinyFloat("0.1").add("0.2").toNumber();
// Or a number:
new TinyFloat(0.1).add(0.2).toNumber();

To convert a TinyFloat instance to a number, use the toNumber method:

new TinyFloat("0.1").add("0.2").toNumber();
//=> 0.3

To convert the TinyFloat instance to a string, use the toString method:

new TinyFloat("0.1").add("0.2").toString();
//=> "0.3000000000000000"

The method pads the number with zeros to the set precision (default is 16).

Both toString and toNumber accept custom precision, effectively rounding the number to the given number of digits after the decimal point:

const tf = new TinyFloat("1").div("1.5");
//=> 0.6666666666666667⁠​​​​​​

tf.toString(2);
//=> "0.67"

tf.toNumber(0);
//=> 1

Changelog

See the changelog.

License

MIT © Sasha Koss