Skip to main content

Getting started

Installation

Kysely can be installed using any of the following package managers:

npm is the default package manager for Node.js, and to where Kysely is published.
Your project is using npm if it has a package-lock.json file in its root folder.

Run the following command in your terminal:

terminal
npm install kysely

Types

For Kysely's type-safety and autocompletion to work, it needs to know your database structure. This requires a TypeScript Database interface, that contains table names as keys and table schema interfaces as values.

Let's define our first database interface:

src/types.ts
import { ColumnType, Generated, Insertable, Selectable, Updateable } from 'kysely'

export interface Database {
person: PersonTable
pet: PetTable
}

export interface PersonTable {
// Columns that are generated by the database should be marked
// using the `Generated` type. This way they are automatically
// made optional in inserts and updates.
id: Generated<number>

first_name: string
gender: 'man' | 'woman' | 'other'

// If the column is nullable in the database, make its type nullable.
// Don't use optional properties. Optionality is always determined
// automatically by Kysely.
last_name: string | null

// You can specify a different type for each operation (select, insert and
// update) using the `ColumnType<SelectType, InsertType, UpdateType>`
// wrapper. Here we define a column `created_at` that is selected as
// a `Date`, can optionally be provided as a `string` in inserts and
// can never be updated:
created_at: ColumnType<Date, string | undefined, never>
}

// You should not use the table schema interfaces directly. Instead, you should
// use the `Selectable`, `Insertable` and `Updateable` wrappers. These wrappers
// make sure that the correct types are used in each operation.
export type Person = Selectable<PersonTable>
export type NewPerson = Insertable<PersonTable>
export type PersonUpdate = Updateable<PersonTable>

export interface PetTable {
id: Generated<number>
name: string
owner_id: number
species: 'dog' | 'cat'
}

export type Pet = Selectable<PetTable>
export type NewPet = Insertable<PetTable>
export type PetUpdate = Updateable<PetTable>
Codegen

For production apps, it is recommended to automatically generate your Database interface by introspecting your production database or Prisma schemas. Generated types might differ in naming convention, internal order, etc. Find out more at "Generating types".

Dialects

For Kysely's query compilation and execution to work, it needs to understand your database's SQL specification and how to communicate with it. This requires a Dialect implementation.

There are 3 built-in Node.js dialects for PostgreSQL, MySQL and SQLite. Additionally, the community has implemented several dialects to choose from. Find out more at "Dialects".

Driver installation

A Dialect implementation usually requires a database driver library as a peer dependency. Let's install it using the same package manager command from before:

Kysely's built-in PostgreSQL dialect uses the "pg" driver library under the hood. Please refer to its official documentation for configuration options.

Run the following command in your terminal:

terminal
npm install pg

I use a different package manager (not npm)

Driverless
Kysely can also work in compile-only mode that doesn't require a database driver. Find out more at "Splitting build, compile and execute code".

Instantiation

Let's create a Kysely instance using the built-in PostgresDialect dialect:

src/database.ts
import { Database } from './types.ts' // this is the Database interface we defined earlier
import { Pool } from 'pg'
import { Kysely, PostgresDialect } from 'kysely'

const dialect = new PostgresDialect({
pool: new Pool({
database: 'test',
host: 'localhost',
user: 'admin',
port: 5434,
max: 10,
})
})

// Database interface is passed to Kysely's constructor, and from now on, Kysely
// knows your database structure.
// Dialect is passed to Kysely's constructor, and from now on, Kysely knows how
// to communicate with your database.
export const db = new Kysely<Database>({
dialect,
})

I use a different package manager (not npm)

I use a different database (not PostgreSQL)

Singleton
In most cases, you should only create a single Kysely instance per database. Most dialects use a connection pool internally, or no connections at all, so there's no need to create a new instance for each request.
keeping secrets
Use a secrets manager, environment variables (DO NOT commit `.env` files to your repository), or a similar solution, to avoid hardcoding database credentials in your code.
kill it with fire
When needed, you can dispose of the Kysely instance, release resources and close all connections by invoking the db.destroy() function.

Querying

Let's implement the person repository:

src/PersonRepository.ts
import { db } from './database'
import { PersonUpdate, Person, NewPerson } from './types'

export async function findPersonById(id: number) {
return await db.selectFrom('person')
.where('id', '=', id)
.selectAll()
.executeTakeFirst()
}

export async function findPeople(criteria: Partial<Person>) {
let query = db.selectFrom('person')

if (criteria.id) {
query = query.where('id', '=', criteria.id) // Kysely is immutable, you must re-assign!
}

if (criteria.first_name) {
query = query.where('first_name', '=', criteria.first_name)
}

if (criteria.last_name !== undefined) {
query = query.where(
'last_name',
criteria.last_name === null ? 'is' : '=',
criteria.last_name
)
}

if (criteria.gender) {
query = query.where('gender', '=', criteria.gender)
}

if (criteria.created_at) {
query = query.where('created_at', '=', criteria.created_at)
}

return await query.selectAll().execute()
}

export async function updatePerson(id: number, updateWith: PersonUpdate) {
await db.updateTable('person').set(updateWith).where('id', '=', id).execute()
}

export async function createPerson(person: NewPerson) {
return await db.insertInto('person')
.values(person)
.returningAll()
.executeTakeFirstOrThrow()
}

export async function deletePerson(id: number) {
return await db.deleteFrom('person').where('id', '=', id)
.returningAll()
.executeTakeFirst()
}

I use a different database (not PostgreSQL)

But wait, there's more!
This is a simplified example with basic CRUD operations. Kysely supports many more SQL features including: joins, subqueries, complex boolean logic, set operations, CTEs, functions (aggregate and window functions included), raw SQL, transactions, DDL queries, etc.
Find out more at Examples.

Summary

We've seen how to install and instantiate Kysely, its dialects and underlying drivers. We've also seen how to use Kysely to query a database.

Let's put it all to the test:

src/PersonRepository.spec.ts
import { sql } from 'kysely'
import { db } from './database'
import * as PersonRepository from './PersonRepository'

describe('PersonRepository', () => {
before(async () => {
await db.schema.createTable('person')
.addColumn('id', 'serial', (cb) => cb.primaryKey())
.addColumn('first_name', 'varchar', (cb) => cb.notNull())
.addColumn('last_name', 'varchar')
.addColumn('gender', 'varchar(50)', (cb) => cb.notNull())
.addColumn('created_at', 'timestamp', (cb) =>
cb.notNull().defaultTo(sql`now()`)
)
.execute()
})

afterEach(async () => {
await sql`truncate table ${sql.table('person')}`.execute(db)
})

after(async () => {
await db.schema.dropTable('person').execute()
})

it('should find a person with a given id', async () => {
await PersonRepository.findPersonById(123)
})

it('should find all people named Arnold', async () => {
await PersonRepository.findPeople({ first_name: 'Arnold' })
})

it('should update gender of a person with a given id', async () => {
await PersonRepository.updatePerson(123, { gender: 'woman' })
})

it('should create a person', async () => {
await PersonRepository.createPerson({
first_name: 'Jennifer',
last_name: 'Aniston',
gender: 'woman',
})
})

it('should delete a person with a given id', async () => {
await PersonRepository.deletePerson(123)
})
})

I use a different database (not PostgreSQL)

Migrations
As you can see, Kysely supports DDL queries. It also supports classic "up/down" migrations. Find out more at Migrations.