How to use Sequelize: install, define models and query a database

What Sequelize is and when to use it

Sequelize is a promise-based ORM tool for Node.js. The README lists support for PostgreSQL, MySQL, MariaDB, SQLite, DB2, Microsoft SQL Server, Snowflake, Oracle DB and Db2 for IBM i. It features transaction support, relations, eager and lazy loading, and read replication.

It suits Node.js backend developers who use a relational SQL database, TypeScript teams that want typed models, and projects that may switch between PostgreSQL, MySQL, SQLite or MSSQL. It is built for relational databases, so it is not the right tool if your project does not use SQL.

• Good fit: Node.js apps on a SQL database that want transactions and relations without raw SQL for every query.
• Poor fit: projects with no SQL database, or developers who prefer writing raw SQL with minimal abstraction.

Install Sequelize with npm

Sequelize is installed through a package manager. You need Node.js and a supported SQL database such as PostgreSQL, MySQL, SQLite or MSSQL. The core package is published as @sequelize/core.

Run the install command from your project folder.

Check the version before you start

The README points to two getting started guides: one for Sequelize 6 (stable) and one for Sequelize 7 (alpha). It also says the project is actively working on its next major version. If you need a stable release for production right now, follow the version 6 guide and check the upgrade notes before moving to a newer major version.

Upgrade notes are linked from the README for moving from v5 to v6 and from v6 to v7.

• Use the Sequelize 6 guide for a stable setup.
• Use the Sequelize 7 guide only if you are comfortable with an alpha release.

Connect to your database and define models

Sequelize maps database tables to JavaScript models. Create a connection to your database, then define the models your app needs. The official getting started guide on sequelize.org walks through the exact connection and model syntax for each release, so follow it for the version you installed.

Because Sequelize supports several dialects, the connection step is where you name the database type you are using. Keeping this in one place makes it easier to switch between PostgreSQL, MySQL, SQLite or MSSQL later.

• Choose the database dialect your project uses.
• Define a model for each table you want to read and write.
• Keep connection settings out of source code and load them from configuration.

Read and write data with models

Once models are defined, you read and write rows through them instead of writing SQL strings. Sequelize's methods are promise-based, so you can use async and await in the usual Node.js style.

For relations, Sequelize supports associations between models, and you can load related data with eager loading or fetch it later with lazy loading.

• Use model methods to create, read, update and delete rows.
• Use eager loading when you need related rows in one query.
• Use lazy loading when related data is only needed sometimes.

Use transactions for multi-step changes

Sequelize has transaction support. Use a transaction when several database writes must succeed or fail together. The exact transaction syntax is in the official documentation for your version, so check it there before copying code.

Get help and contribute

The project lists resources including its documentation, changelog, GitHub discussions, Slack and Stack Overflow. A community CLI and several tools for GraphQL, CockroachDB and YugabyteDB are also linked from the README.

The project also says it is looking for new maintainers to help release the next major version. Contribution guidelines are in CONTRIBUTING.md in the repository.

• Documentation and quick start: sequelize.org
• Questions: GitHub discussions or Stack Overflow under sequelize.js
• Contributing: read CONTRIBUTING.md