Selecting a Data Modeling Library for Node.js Services
bookshelf, mongoose, sequelize, and waterline are libraries that help Node.js applications interact with databases by organizing data into models. mongoose is designed for MongoDB and uses a document-based approach, while sequelize, bookshelf, and waterline support relational SQL databases like PostgreSQL and MySQL. These tools handle connections, define data shapes, and simplify queries so developers do not need to write raw SQL or database commands for every action. Choosing the right one depends on your database type, project size, and long-term maintenance needs.
Npm Package Weekly Downloads Trend
Github Stars Ranking
Stat Detail
Bookshelf vs Mongoose vs Sequelize vs Waterline: Data Modeling Compared
When building Node.js backends β whether for API servers, serverless functions, or full-stack frameworks β you need a reliable way to talk to your database. bookshelf, mongoose, sequelize, and waterline all solve this problem, but they target different databases and work in different ways. Let's compare how they handle common engineering tasks.
ποΈ Database Support: SQL vs NoSQL
The first decision is your database type. This choice often dictates which library you can use.
mongoose works only with MongoDB.
- It treats data as JSON-like documents.
- Great for flexible schemas and hierarchical data.
// mongoose: MongoDB only
const mongoose = require('mongoose');
// Connects to MongoDB instance
mongoose.connect('mongodb://localhost:27017/myapp');
sequelize works with SQL databases.
- Supports PostgreSQL, MySQL, MariaDB, SQLite, and SQL Server.
- Enforces strict table structures and relationships.
// sequelize: SQL databases
const { Sequelize } = require('sequelize');
// Connects to SQL instance
const sequelize = new Sequelize('postgres://user:pass@localhost:5432/db');
bookshelf works with SQL databases via Knex.js.
- Supports the same SQL dialects as Knex (Postgres, MySQL, etc.).
- Focuses on mapping rows to models.
// bookshelf: SQL databases via Knex
const knex = require('knex')({ client: 'pg' });
const bookshelf = require('bookshelf')(knex);
// Uses Knex connection for SQL
waterline works with both SQL and NoSQL via adapters.
- Can switch between Mongo, Postgres, MySQL, etc., by changing config.
- Historically tied to Sails.js but works standalone.
// waterline: Adapter-based
const Waterline = require('waterline');
const waterline = new Waterline();
// Adapters determine DB type (mongo, postgres, etc.)
π Defining Data Shapes: Schemas and Models
Each library has a different way to define what your data looks like.
mongoose uses Schemas to define document structure.
- You define types and validation rules upfront.
- Flexible but enforces consistency.
// mongoose: Schema definition
const userSchema = new mongoose.Schema({
name: String,
email: { type: String, unique: true }
});
const User = mongoose.model('User', userSchema);
sequelize uses Model Definitions with data types.
- You define columns and types explicitly.
- Supports strong typing and validation.
// sequelize: Model definition
const User = sequelize.define('User', {
name: { type: Sequelize.STRING },
email: { type: Sequelize.STRING, unique: true }
});
bookshelf extends a Base Model with table names.
- Less focus on types, more on mapping rows.
- Relies on Knex for schema management.
// bookshelf: Model extension
const User = bookshelf.Model.extend({
tableName: 'users',
// Validation handled externally or via plugins
});
waterline uses Collection Definitions.
- Defines attributes and types similar to ORM models.
- Adapter handles the underlying storage details.
// waterline: Collection definition
const User = waterline.Collection.extend({
tableName: 'users',
attributes: {
name: { type: 'string' },
email: { type: 'string', unique: true }
}
});
π Fetching Data: Querying Patterns
How you retrieve data varies significantly between these tools.
mongoose uses Chainable Queries on models.
- Methods like
find,findOne,findByIdare standard. - Returns plain JavaScript objects or documents.
// mongoose: Finding documents
const users = await User.find({ role: 'admin' });
const user = await User.findOne({ email: 'test@example.com' });
sequelize uses Finder Methods with options objects.
findAll,findOne,findByPkare common.- Supports complex includes for relationships.
// sequelize: Finding rows
const users = await User.findAll({ where: { role: 'admin' } });
const user = await User.findOne({ where: { email: 'test@example.com' } });
bookshelf uses Knex-style Querying on models.
- Uses
where,fetch,fetchAllmethods. - Closely mirrors SQL logic.
// bookshelf: Fetching models
const users = await User.where({ role: 'admin' }).fetchAll();
const user = await User.where({ email: 'test@example.com' }).fetch();
waterline uses Criteria Objects for queries.
find,findOneaccept simple objects.- Abstracts away SQL vs NoSQL query differences.
// waterline: Finding records
const users = await User.find({ role: 'admin' });
const user = await User.findOne({ email: 'test@example.com' });
π οΈ Maintenance and Ecosystem Health
For professional developers, long-term support is critical. You do not want to build on a library that stops receiving updates.
mongoose is highly active.
- Regular releases and security patches.
- Massive ecosystem of plugins and tutorials.
- Safe choice for new MongoDB projects.
sequelize is highly active.
- Consistent updates for SQL features and security.
- Strong TypeScript support and community.
- Safe choice for new SQL projects.
bookshelf has low activity.
- Development has slowed significantly in recent years.
- Still works, but new features are rare.
- Consider for legacy maintenance, not new builds.
waterline has low activity standalone.
- Primarily maintained for Sails.js framework users.
- Standalone usage is less common and harder to support.
- Evaluate carefully if not using Sails.
π Summary: Key Differences
| Feature | mongoose | sequelize | bookshelf | waterline |
|---|---|---|---|---|
| Database | π MongoDB only | ποΈ SQL (Postgres, MySQL) | ποΈ SQL (via Knex) | π Both (via adapters) |
| Style | π Document ODM | ποΈ Relational ORM | ποΈ Relational ORM | π Datastore Agnostic |
| Activity | π’ High | π’ High | π‘ Low | π‘ Low |
| Best For | π New Mongo Apps | π’ Enterprise SQL | π°οΈ Legacy Knex Apps | β΅ Sails.js Apps |
π‘ The Big Picture
mongoose and sequelize are the industry standards for new projects. Choose mongoose for MongoDB and sequelize for SQL databases. They have the best documentation, community support, and active maintenance.
bookshelf and waterline are older tools with specific use cases. bookshelf fits if you love Knex and need a light model layer. waterline fits if you are already using Sails.js. For most other scenarios, the active maintenance of mongoose and sequelize makes them the safer choice for long-term stability.
Final Thought: Your database choice should drive your library choice. Pick your database first based on data needs, then pick the library that supports it best with active community backing.
How to Choose: mongoose vs bookshelf vs sequelize vs waterline
- mongoose:
Choose
mongooseif your application uses MongoDB and you need a robust schema system with built-in validation. It is the standard choice for Node.js and MongoDB, offering strong community support and extensive plugins. This is ideal for content-heavy apps, real-time data, or when your data structure changes frequently. - bookshelf:
Choose
bookshelfif you are already using Knex.js and want a thin model layer on top of it for SQL databases. It is suitable for legacy projects or teams that prefer explicit control over SQL queries without heavy abstraction. However, be aware that development activity has slowed, so it may not be the best fit for new greenfield projects requiring long-term support. - sequelize:
Choose
sequelizeif you are working with relational databases like PostgreSQL, MySQL, or SQLite and want a feature-rich ORM. It supports migrations, associations, and transactions out of the box, making it great for complex business logic. It is widely maintained and a safe bet for enterprise-grade SQL applications. - waterline:
Choose
waterlineprimarily if you are building an application with the Sails.js framework, as it is tightly integrated into that ecosystem. It supports both SQL and NoSQL databases through adapters, offering flexibility if you need to switch database types later. For standalone Node.js projects outside of Sails, other options may offer better documentation and community support.
Popular Comparisons
README for mongoose
Mongoose
Mongoose is a MongoDB object modeling tool designed to work in an asynchronous environment. Mongoose supports Node.js and Deno (alpha).
Documentation
The official documentation website is mongoosejs.com.
Mongoose 9.0.0 was released on November 21, 2025. You can find more details on backwards breaking changes in 9.0.0 on our docs site.
Support
Plugins
Check out the plugins search site to see hundreds of related modules from the community. Next, learn how to write your own plugin from the docs or this blog post.
Contributors
Pull requests are always welcome! Please base pull requests against the master
branch and follow the contributing guide.
If your pull requests makes documentation changes, please do not
modify any .html files. The .html files are compiled code, so please make
your changes in docs/*.pug, lib/*.js, or test/docs/*.js.
View all 400+ contributors.
Installation
First install Node.js and MongoDB, then install the mongoose package using your preferred package manager:
Using npm
npm install mongoose
Using pnpm
pnpm add mongoose
Using Yarn
yarn add mongoose
Using Bun
bun add mongoose
Mongoose 6.8.0 also includes alpha support for Deno.
Importing
// Using Node.js `require()`
const mongoose = require('mongoose');
// Using ES6 imports
import mongoose from 'mongoose';
Or, using Deno's createRequire() for CommonJS support as follows.
import { createRequire } from 'https://deno.land/std@0.177.0/node/module.ts';
const require = createRequire(import.meta.url);
const mongoose = require('mongoose');
mongoose.connect('mongodb://127.0.0.1:27017/test')
.then(() => console.log('Connected!'));
You can then run the above script using the following.
deno run --allow-net --allow-read --allow-sys --allow-env mongoose-test.js
Mongoose for Enterprise
Available as part of the Tidelift Subscription
The maintainers of mongoose and thousands of other packages are working with Tidelift to deliver commercial support and maintenance for the open source dependencies you use to build your applications. Save time, reduce risk, and improve code health, while paying the maintainers of the exact dependencies you use. Learn more.
Overview
Connecting to MongoDB
First, we need to define a connection. If your app uses only one database, you should use mongoose.connect. If you need to create additional connections, use mongoose.createConnection.
Both connect and createConnection take a mongodb:// URI, or the parameters host, database, port, options.
await mongoose.connect('mongodb://127.0.0.1/my_database');
Once connected, the open event is fired on the Connection instance. If you're using mongoose.connect, the Connection is mongoose.connection. Otherwise, mongoose.createConnection return value is a Connection.
Note: If the local connection fails then try using 127.0.0.1 instead of localhost. Sometimes issues may arise when the local hostname has been changed.
Important! Mongoose buffers all the commands until it's connected to the database. This means that you don't have to wait until it connects to MongoDB in order to define models, run queries, etc.
Defining a Model
Models are defined through the Schema interface.
const Schema = mongoose.Schema;
const ObjectId = Schema.ObjectId;
const BlogPost = new Schema({
author: ObjectId,
title: String,
body: String,
date: Date
});
Aside from defining the structure of your documents and the types of data you're storing, a Schema handles the definition of:
- Validators (async and sync)
- Defaults
- Getters
- Setters
- Indexes
- Middleware
- Methods definition
- Statics definition
- Plugins
- pseudo-JOINs
The following example shows some of these features:
const Comment = new Schema({
name: { type: String, default: 'hahaha' },
age: { type: Number, min: 18, index: true },
bio: { type: String, match: /[a-z]/ },
date: { type: Date, default: Date.now },
buff: Buffer
});
// a setter
Comment.path('name').set(function(v) {
return capitalize(v);
});
// middleware
Comment.pre('save', function(next) {
notify(this.get('email'));
next();
});
Take a look at the example in examples/schema/schema.js for an end-to-end example of a typical setup.
Accessing a Model
Once we define a model through mongoose.model('ModelName', mySchema), we can access it through the same function
const MyModel = mongoose.model('ModelName');
Or just do it all at once
const MyModel = mongoose.model('ModelName', mySchema);
The first argument is the singular name of the collection your model is for. Mongoose automatically looks for the plural version of your model name. For example, if you use
const MyModel = mongoose.model('Ticket', mySchema);
Then MyModel will use the tickets collection, not the ticket collection. For more details read the model docs.
Once we have our model, we can then instantiate it, and save it:
const instance = new MyModel();
instance.my.key = 'hello';
await instance.save();
Or we can find documents from the same collection
await MyModel.find({});
You can also findOne, findById, update, etc.
const instance = await MyModel.findOne({ /* ... */ });
console.log(instance.my.key); // 'hello'
For more details check out the docs.
Important! If you opened a separate connection using mongoose.createConnection() but attempt to access the model through mongoose.model('ModelName') it will not work as expected since it is not hooked up to an active db connection. In this case access your model through the connection you created:
const conn = mongoose.createConnection('your connection string');
const MyModel = conn.model('ModelName', schema);
const m = new MyModel();
await m.save(); // works
vs
const conn = mongoose.createConnection('your connection string');
const MyModel = mongoose.model('ModelName', schema);
const m = new MyModel();
await m.save(); // does not work b/c the default connection object was never connected
Embedded Documents
In the first example snippet, we defined a key in the Schema that looks like:
comments: [Comment]
Where Comment is a Schema we created. This means that creating embedded documents is as simple as:
// retrieve my model
const BlogPost = mongoose.model('BlogPost');
// create a blog post
const post = new BlogPost();
// create a comment
post.comments.push({ title: 'My comment' });
await post.save();
The same goes for removing them:
const post = await BlogPost.findById(myId);
post.comments[0].deleteOne();
await post.save();
Embedded documents enjoy all the same features as your models. Defaults, validators, middleware.
Middleware
See the docs page.
Intercepting and mutating method arguments
You can intercept method arguments via middleware.
For example, this would allow you to broadcast changes about your Documents every time someone sets a path in your Document to a new value:
schema.pre('set', function(next, path, val, typel) {
// `this` is the current Document
this.emit('set', path, val);
// Pass control to the next pre
next();
});
Moreover, you can mutate the incoming method arguments so that subsequent middleware see different values for those arguments. To do so, just pass the new values to next:
schema.pre(method, function firstPre(next, methodArg1, methodArg2) {
// Mutate methodArg1
next('altered-' + methodArg1.toString(), methodArg2);
});
// pre declaration is chainable
schema.pre(method, function secondPre(next, methodArg1, methodArg2) {
console.log(methodArg1);
// => 'altered-originalValOfMethodArg1'
console.log(methodArg2);
// => 'originalValOfMethodArg2'
// Passing no arguments to `next` automatically passes along the current argument values
// i.e., the following `next()` is equivalent to `next(methodArg1, methodArg2)`
// and also equivalent to, with the example method arg
// values, `next('altered-originalValOfMethodArg1', 'originalValOfMethodArg2')`
next();
});
Schema gotcha
type, when used in a schema has special meaning within Mongoose. If your schema requires using type as a nested property you must use object notation:
new Schema({
broken: { type: Boolean },
asset: {
name: String,
type: String // uh oh, it broke. asset will be interpreted as String
}
});
new Schema({
works: { type: Boolean },
asset: {
name: String,
type: { type: String } // works. asset is an object with a type property
}
});
Driver Access
Mongoose is built on top of the official MongoDB Node.js driver. Each mongoose model keeps a reference to a native MongoDB driver collection. The collection object can be accessed using YourModel.collection. However, using the collection object directly bypasses all mongoose features, including hooks, validation, etc. The one
notable exception that YourModel.collection still buffers
commands. As such, YourModel.collection.find() will not
return a cursor.
API Docs
Mongoose API documentation, generated using dox and acquit.
Related Projects
MongoDB Runners
Unofficial CLIs
Data Seeding
Express Session Stores
License
Copyright (c) 2010 LearnBoost <dev@learnboost.com>
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the 'Software'), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.