npm i warehousejs


Warehouse.js is a data storage layer for node.js and browser providing unified API for many supported storage engines (MongoDB, MySQL, SQLite, in-memory and others).

by Jakub Dundalek

0.3.0 (see all)TypeScript:Not Found
npm i warehousejs


Warehouse.js is a data storage layer for node.js and browser providing unified API for many supported storage engines (MongoDB, MySQL, SQLite, in-memory and others). You can create REST server or client with only few lines of code.

  • Supports multiple backends (MongoDB, MySQL, SQLite, in-memory and others)
  • Create REST server using few lines of code (compatible with Backbone.js)
  • Unified API for CRUD operations
  • Powerful query language RQL
  • MIT License

Make a quick protoype of your application without using an external database. Then switch to a robust storage (like MongoDB or MySQL) without the need to change your code.


npm install git://github.com/dundalek/warehouse.git

REST server with express in node.js

var express = require('express'),
    warehouse = require('warehouse'),
    MongoBackend = require('warehouse/backend/mongodb');

var app = express.createServer(),
    store = new MongoBackend().objectStore('item');

warehouse.applyRoutes(app, store);


The example server above can be used with Backbone.js.

// client

var Model = Backbone.Model.extend({
   idAttribute: '_id' // Mongo uses _id for primary key

var items = new Backbone.Collection.extend({
   url: '/item',
   model: Model


var john = items.create({name: 'John', age: 25});

john.set('age', 30);



The API is similar to IndexedDB API (add, get, put, delete). It is based on asynchronous promises (using Q library).

var jack = {id: 15, firstname: 'Jack', lastname: 'Hammer', age: 35};

// Create
     .then(function(result) { /* success */ })
     .fail(function(error) { /* error */ });

// Read
     .then(function(result) { console.log(result.name) }); // outputs: Jack

// Update
jack.age = 40;

// Delete

// Delete all


Queries are implemented using RQL.

// get items with id=15
     .then(function(result) {}); // result is an Array

// get items with age >= 21

// get items with price < 100, sort by acending price and descending rating



Backend represents DB that has collections for storing items. Common methods are:

  • objectStoreNames () - list collections or tables
  • objectStore(String name, [Object options]) - gets object store with specified name. If it does not exists then it is created.
  • createObjectStore (String name, [Object options]) - creates object store with specified name
  • deleteObjectStore (String name) - deletes object store with specified name (equivalent for droping table in SQL)
  • open () - opens database connection. You don't have to worry about it most of the times, because connection will be automatically opened first time when it is needed (lazy loading).
  • close () - closes database connection
  • isClosed () - is the connection closed?

Options for stores are:

  • keyPath - attribute that represents primary key (default 'id').


Store represents collection of items (e.g. collection for MongoDB or table for SQL). Common methods are:

  • put (value, [key | directives])
  • add (value, [key | directives])
  • delete (key | directives)
  • get (key | directives)
  • clear ()
  • query (String query)

Key is a built-in type - String, Number, ... Directives is a object e.g. {key: '123'}

Common attributes:

  • String name
  • String keyPath



Backend for MongoDB using node-mongodb-native.

var options = {
    // default connection options
    host: 'localhost',
    port: 27017,
    database: 'default',

    // optional authentication credentials
    user: 'user',
    password: 'pass'

var MongoBackend = require('warehouse/backend/mongodb'),
    backend = new MongoBackend(options),
    store = backend.objectStore('item');

You can use db and collection methods to get native driver objects.

store.db().then(function(db) {
    // native driver method
    db.ensureIndex('item', {price: 1}, function(err, result) {});

store.collection().then(function(collection) {
    // native collection method
    collection.find({tags: {$size: 2}}, function(err, result) {});


Backend for SQL databases, supports MySQL and SQLite. The underlying library node-persist also supports PostgreSQL and Oracle. However, those databases were not tested.

// options when using MySQL
var options = {
    driver: 'mysql',

    // default connection options
    host: 'localhost',
    port: 3306,
    name: 'default', // database name

    // optional authentication credentials
    user: 'user',
    password: 'pass'

// options when using SQLite
var options = {
    driver: 'sqlite3',
    filename: '/path/to/file.db',

var SqlBackend = require('warehouse/backend/sql'),
    backend = new SqlBackend(options),
    store = backend.objectStore('item');

If you need to execute advanced queries, runSql and runSqlAll methods can be used.

// runSql(sql, values) - runs a sql statement that does not return results (INSERT, UPDATE, etc).
store.runSql("UPDATE people SET age = ?", [32])
     .then(function(status) {
         // people updated

// runSqlAll(sql, values) - runs a sql statement that returns results (ie SELECT).
store.runSqlAll("SELECT * FROM people WHERE age = ?", [32])
     .then(function(items) {
         // items is an array containing all the people with age 32


Backend using ElasticSearch.

This backend works under both server and browser.

Pass url option to specify remote server address.

var ElasticSearchBackend = require('warehouse/backend/elasticsearch'),
    backend = new ElasticSearchBackend({url: 'http://example.com/index'}),
    store = backend.objectStore('item');


Store items in memory using native Arrays and Objects. It is useful for making quick prototypes without the need of external dependencies.

This backend works under both server and browser.

Store takes optional argument json, which loads initial data into the datastore. Argument can be one of

  • Array
  • Object
  • Array JSON-encoded in String
  • Object JSON-encoded in String
  • String url of an remote resource, starts with 'http'

You can alternatively use fromJson and getJSON

var MemoryBackend = require('warehouse/backend/memory'),
    backend = new MemoryBackend(),
    store, data;

// Array
data = [
  {id: 1, name: 'John'},
  {id: 2, name: 'Sarah'}

// Array in JSON
data = '[{"id":1,"name":"John"},{"id":2,"name":"Sarah"}]';

// Object
data = {
  1: {id: 1, name: 'John'},
  2: {id: 2, name: 'Sarah'}

// Object in JSON
data = {"1":{"id":1,"name":"John"},"2":{"id":2,"name":"Sarah"}};

// create the store with initial data
store = backend.objectStore('item', {json: data});

// create the store and set data laterr
store = backend.objectStore('item');


This backend is using remote REST server for storage. It is useful to access remote services.

If you combine it with a server (acting as a transparent proxy), you can basically use all implemented backends as if they were available for client.

This backend works under both server and browser.

Pass url option to specify remote server address (default is '/').

var RestBackend = require('warehouse/backend/rest'),
    backend = new RestBackend({url: 'http://example.com'}),
    store = backend.objectStore('item');

Local Storage

Implements storage using W3C Web Storage (also known as DOM Storage or Local Storage).

Names of Object Stores have a special meaning:

  • session is a storage that is available for the duration of the page session
  • local is persistent (same-origin rules are applied}

This backend works only in browser.

var LocalBackend = require('warehouse/backend/local'),
    backend = new LocalBackend();

// session only
var sessionStore = backend.objectStore('session');

// persistent
var localStore = backend.objectStore('local');