SQL you know but type safe parameterized, expressive & portable

litdb contains LINQ-like type-safe SQL builders for TypeScript/JavaScript for writing type-safe expressive SQL that's parameterized & portable across SQLite, MySQL & PostgreSQL

SQL-like

No need to learn a new query language, litdb lets you write familiar SQL you know using its type-safe expressions and query builders that's safe by default.

APIs are designed to keep a close affinity to SQL making it clear what SQL is being generated, expressions utilize tagged templates to take away the tedium of maintaining table and column references, parameterizing values and catering for different SQL dialects to retain the expressive freedom of SQL, but portable across multiple RDBMS's.

The primary difference between SQL is how queries are constructed, where the SELECT Query Builder wants queries to be constructed in the order they're run, by specifying the data sources first, i.e. the FROM table first, followed by any JOINs (just like LINQ) which allows litdb query builders to provide type safety and intellisense for the rest of the query.

kb

Small

  • 9kb .min + .zip, 36kb .min
  • 0 dependencies

Install

To use litdb with your favorite ORM, no driver is required. Just use the litdb package directly:

npm install litdb

litdb is also available as a module, where it can be used directly in the browser:

<script type="module">
import { sqlite as $ } from "https://unpkg.com/litdb/dist/index.min.js"
    
const { sql, params } = $.from(Contact).select(c => $`${c.name}`).build()
</script>

To get the most out of litdb we recommend using text editors that supports TypeScript definitions (e.g. VS Code, JetBrains IDEs, neovim, etc.)

LitDB Drivers

Lightweight drivers with native support for its typed SQL Builders and parameterized SQL Expressions are also available for the popular databases:

Bun SQLite

Use with Bun's native SQLite3 driver:

bun install @litdb/bun-sqlite

Node better-sqlite

Use with Node better-sqlite3:

npm install @litdb/better-sqlite

PostgreSQL

Use with the postgres.js client:

npm install @litdb/postgres

MySQL

Use with the mysql2 client:

npm install @litdb/mysql2

Request a Driver

If you'd like to see a driver for a specific client, please open or vote for a feature request on litdb's GitHub Discussions.

Driver Usage

litdb drivers are lightweight data adapters providing a number of convenience APIs for executing SQL and parameters. They can be used with our without litdb SQL Builders, but offer the most value when used together.

The same APIs are available across all drivers, so you can easily switch between them. They include both sync APIs recommended for SQLite libraries that use SQLite's native blocking APIs, whilst async APIs should be used for all other databases, e.g. PostgreSQL and MySQL.

Example of using the Bun SQLite driver:

db.ts

import { connect } from "@litdb/bun-sqlite"

export const connection = connect("app.db") // WAL enabled by default
export const { $, sync:db, async, native } = connection

TIP

When needed use native to access underlying driver (e.g. bun:sqlite Database)

app.ts

import { $, db } from "./db"
import { Contact } from "./models"

db.dropTable(Contact)
db.createTable(Contact)
db.insertAll([
    new Contact({ name:"John Doe", email:"[email protected]" }),
    new Contact({ name:"Jane Doe", email:"[email protected]" }),
])

const janeEmail = '[email protected]'
const jane = db.one<Contact>($.from(Contact).where(c => $`${c.email} = ${janeEmail}`))!

// Insert examples
const { lastInsertRowid: bobId } = db.insert(new Contact({ name:"Bob", email:"[email protected]" }))
const { lastInsertRowid } = db.exec`INSERT INTO Contact(name,email) VALUES ('Jo','[email protected]')`
const name = 'Alice', email = '[email protected]'
db.exec`INSERT INTO Contact(name,email) VALUES (${name}, ${email})`

// Typed SQL fragment with named param example
const hasId = <Table extends { id:number }>(id:number|bigint) =>
    (x:Table) => $.sql($`${x.id} = $id`, { id })

const contacts = db.all($.from(Contact).into(Contact))                // => Contact[]
const bob = db.one($.from(Contact).where(hasId(bobId)).into(Contact)) // => Contact    
const contactsCount = db.value($.from(Contact).select`COUNT(*)`)      // => number
const emails = db.column($.from(Contact).select(c => $`${c.email}`))  // => string[]
const contactsArray = db.arrays($.from(Contact))                      // => any[][]
const bobArray = db.array($.from(Contact).where(hasId(bobId)))        // => any[]

// Update examples
jane.email = '[email protected]'
db.update(jane)                           // Update all properties
db.update(jane, { onlyProps:['email'] })  // Update only email
db.exec($.update(Contact).set({ email:jane.email }).where(hasId(jane.id))) // query builder

// Delete examples
db.delete(jane)
db.exec($.deleteFrom(Contact).where(hasId(jane.id))) // query builder

Same source is compatible with other sync drivers, just replace @litdb/bun-sqlite with @litdb/better-sqlite to use with litdb/better-sqlite.

Async Driver Usage

The same APIs (returning promises) are available for async access to remote RDBMS's like PostgreSQL and MySQL.

Example of using the postgres.js driver:

db.ts

import { connect } from "@litdb/postgres"

export const connection = connect({ hostname, database, user, password })
export const { $, async:db, native:sql } = connection

TIP

When needed use sql to access the native postgres.js sql function

app.ts

import { $, db } from "./db"
import { Contact } from "./models"

await db.dropTable(Contact)
await db.createTable(Contact)
await db.insertAll([
    new Contact({ name:"John Doe", email:"[email protected]" }),
    new Contact({ name:"Jane Doe", email:"[email protected]" }),
])

const janeEmail = '[email protected]'
const jane = await db.one<Contact>($.from(Contact).where(c => $`${c.email}=${janeEmail}`))

// Insert examples
const { lastInsertRowid:bobId } = await db.insert(
    new Contact({ name:"Bob", email:"[email protected]"}))

const { lastInsertRowid } = await db.exec
    `INSERT INTO Contact(name,email) VALUES('Jo','[email protected]')`

const name = 'Alice', email = '[email protected]'
await db.exec`INSERT INTO Contact(name,email) VALUES (${name}, ${email})`

// Typed SQL fragment with named param example
const hasId = <Table extends { id:number }>(id:number|bigint) =>
    (x:Table) => $.sql($`${x.id} = $id`, { id })

const contacts = await db.all($.from(Contact).into(Contact))                // => Contact[]
const bob = await db.one($.from(Contact).where(hasId(bobId)).into(Contact)) // => Contact
const contactsCount = await db.value($.from(Contact).rowCount())            // => number
const emails = await db.column($.from(Contact).select(c => $`${c.email}`))  // => string[]
const contactsArray = await db.arrays($.from(Contact))                      // => any[][]
const bobArray = await db.array($.from(Contact).where(hasId(bobId)))        // => any[]

// Update examples
jane.email = '[email protected]'
await db.update(jane)                           // Update all properties
await db.update(jane, { onlyProps:['email'] })  // Update only email
// query builder
await db.exec($.update(Contact).set({ email:jane.email }).where(hasId(jane.id)))

// Delete examples
await db.delete(jane)
await db.exec($.deleteFrom(Contact).where(hasId(jane.id))) // query builder

Same source is compatible with other async drivers, just replace @litdb/postgres with @litdb/mysql2 to use with litdb/mysql2.

Type Safe

Get productive intelli-sense for quickly accessing properties that exist your data model or design-time type-checking and compile-time static analysis errors for those that don't.

Only reference table and columns that are included in your query, quickly identify missing tables, columns and any typos:

Safe Refactoring

All references are statically typed to your models so you can safely refactor with confidence!

Composable

Queries are highly composable where external references can be used across multiple Query Builders and SQL fragments to easily create and compose multiple complex queries with shared references.

SQL Builders and SQL fragments can be embedded inside other query builders utilizing the full expressiveness of SQL where their SQL and parameters are merged into the parent query.

// Use external aliased table references across multiple query builders
const [ c, o ] = [ $.ref(Contact,'c'), $.ref(Order,'o') ]

const now = new Date()
const monthAgo = new Date(now.setDate(now.getDate()-30)).toISOString().split('T')[0]
const last30Days = $.from(Order,'o2')
    .where(o2 => $`${o2.contactId} = ${c.id}`)
    .and(o2 => $`${o2.createdAt} >= ${monthAgo}`)
    .select(o2 => $`COUNT(${o2.id})`)

const recentOrder = $.from(Order,'o3')
    .where(o3 => $`${o3.contactId} = ${c.id}`)
    .select(o3 => $`MAX(${o3.createdAt})`)

// Example of SQL Fragment with parameter
const startOfYear = `2024-01-01`
const o4 = $.ref(Order,'o4')
const totalOrders = $`SELECT SUM(${o4.total}) 
     FROM ${o4} o4 
    WHERE ${o4.contactId} = ${c.id} 
      AND ${o4.createdAt} >= ${startOfYear}`

// Compose queries from multiple query builders and SQL fragments
const q = $.from(c)
    .join(o, { on:(c,o) => $`${c.id} = ${o.contactId}`})
    .where`${o.createdAt} = (${recentOrder})`
    .select`
        ${c.id}, 
        ${c.name}, 
        ${o.createdAt} AS recentOrder, 
        (${last30Days}) AS last30Days,
        (${totalOrders}) AS totalOrders`
    .orderBy`last30Days DESC`

Generated SQL

SELECT c."id", 
       c."name", 
       o."createdAt" AS recentOrder, 
       (SELECT COUNT(o2."id")
          FROM "Order" o2
         WHERE o2."contactId" = c."id"
           AND o2."createdAt" >= $_1) AS last30Days,
       (SELECT SUM(o4."total") 
          FROM "Order" o4
         WHERE o4."contactId" = c."id" 
           AND o4."createdAt" >= $_2) AS totalOrders
 FROM "Contact" c
 JOIN "Order" o ON c."id" = o."contactId"
WHERE o."createdAt" = (SELECT MAX(o3."createdAt")
        FROM "Order" o3
        WHERE o3."contactId" = c."id")
ORDER BY last30Days DESC

PARAMS

_1: 2024-10-12
_2: 2024-01-01