kuuid

kuuid

If you need unique identifiers in your Node.js app for use in a database such as Apache CouchDB or Cloudant, then kuuid can generate them. The ids it generates are:

• uniform - all ids are 32 characters long.
• unique - no two ids will be the same, or at least the likelihood of a clash is vanishingly small.
• time-sortable - the ids sort into time order, with one second precision.

If a kuuid-generated id were used as a database's unique identifier, it would sort roughly in time order.

Installation

Add kuuid to your Node.js project with:

npm install --save kuuid

Import the library into your code with:

let kuuid = require('kuuid')

Generating an id

Simply call the kuuid function to get an id:

let id = kuuid()
// 001fgS7k4gJxqY1aXpni3gHuOy0WusLe

You can use such an id as a unique identifier in your database records:

let doc = {
  _id: kuuid(),
  name: 'Glynn',
  location: 'UK',
  verified: true
}
// {"_id":"001fgS954GN35e4NJPyK1W9aiE44m2xD","name":"Glynn","location":"UK","verified":true}
db.insert(doc)

Supplying no parameter to kuuid() uses the current time to generate the timestamp part of the id. You may also supply your own date/time:

// 'now'
kuuid()

// ISO String 
kuuid('2018-07-20T10:10:34.234Z')

// millseconds since 1970
kuuid(1514764800)

How does it work?

A kuuid string has two parts:

• 8 characters representing the number of seconds since 1st January 1970.
• 24 characters containing random data.

The front eight characters allow the string to be sorted by time. Two ids created in the same second will have the same front eight characters. The remaining 24 characters contain 128 bits of random data.

The strings are encoded in "base 62" (i.e using digits and uppercase/lowercase letters) to pack more information into a smaller space.

Points to note

1. The kuuid library can only be used to store dates after 1970-01-01.
2. The random number is genreated using Node.js's crypto.randomBytes which is a secure, if slow, source of random information.
3. The character set used by the base-62 encoding algorithm differs from other algorithms I've seen to ensure that it sorts correctly in a CouchDB _id field.