Configuring Drizzle kit

Configuration

By default, Drizzle reads the drizzle.config.ts file located in the project’s root folder. If you need to specify a different configuration file, you can utilize the --config=<path> command-line option after any drizzle-kit command.

Drizzle Kit supports multiple formats for the configuration file, including .ts, .js. This flexibility allows you to choose the format that best suits your needs and preferences.

Here are some examples of each file format usage:

Typescript example

import { defineConfig } from 'drizzle-kit/utils'

export default defineConfig({
  schema: "./schema.ts",
  out: "./drizzle"
})

JavaScript example

/** @type { import("drizzle-kit").Config } */
export default {
  schema: "./schema.ts",
  out: "./drizzle"
};

Options

schema

Type: string | string[]
Commands: generate, check, up, push, drop

schema param lets you define where your schema file/files live.

You can have as many separate schema files as you want and define paths to them using glob or array of globs syntax.

Usage

Single file

Multiple files

Glob pattern

export default {
  schema: "./schema.ts",
} satisfies Config;

export default {
  schema: ["./user.sql.ts", "./post.sql.ts"],
} satisfies Config;

export default {
  schema: "./schema/*.ts",
} satisfies Config;

out

Type: string | string[]
Default: drizzle
Commands: generate, check, up, drop, introspect

The out parameter allows you to define the folder for your migrations.

In this folder, drizzle-kit will:

Store migration files, meta, and journal
Output the schema.ts file when running drizzle-kit introspect

Usage

export default {
  out: "./drizzle",
} satisfies Config;

driver

Type: 'pg' | 'mysql2' | 'better-sqlite' | 'libsql' | 'turso' | 'd1'
Commands: push, introspect

⚠️

Current param is available in [email protected] and higher. In previous versions you should not use this param

The driver parameter is responsible for explicitly providing a driver to use when accessing a database for the introspect and push commands. It is also useful for HTTP-based databases where connecting using a database connection url is not possible. For example, Turso and Cloudflare D1!

Usage

export default {
  driver: "d1"
} satisfies Config;

dbCredentials

Type: PostgresCredentials, MySQLCredentials, SQLiteCredentials, TursoCredentials
Commands: push, introspect

💡

Each driver should have dbCredentials field to provide a means of connecting to the database.

PostgresCredentials (pg)

Option 1

Option 2

export default {
  dbCredentials: {
    connectionString: '',
  }
} satisfies Config;

export default {
  dbCredentials: {
    host: "",
    port: "",
    user: "",
    password: "",
    database: "",
    ssl: true,
  }
} satisfies Config;

MySQLCredentials (mysql2)

Option 1

Option 2

export default {
  dbCredentials: {
    uri: '',
  }
} satisfies Config;

export default {
  dbCredentials: {
    host: "",
    port: "",
    user: "",
    password: "",
    database: "",
  }
} satisfies Config;

SQLiteCredentials (libsql, better-sqlite3)

export default {
  dbCredentials: {
    url: '', // 👈 this could also be a path to the local sqlite file
  }
} satisfies Config;

TursoCredentials (turso)

export default {
  dbCredentials: {
    url: '',
    authToken: ''
  }
} satisfies Config;

D1Credentials (d1)

export default {
  dbCredentials: {
    wranglerConfigPath: 'wrangler.toml',
    dbName: "d1-test"
  }
} satisfies Config;

breakpoints

Type: boolean
Default: true (starting from [email protected])
Commands: generate, introspect

breakpoints param lets you enable/disable SQL statement breakpoints in generated migrations. It’s optional and true by default, it’s necessary to properly apply migrations on databases, that do not support multiple DDL alternation statements in one transaction(MySQL, SQLite) and Drizzle ORM has to apply them sequentially one by one.

CREATE TABLE `book` (
  `id` integer PRIMARY KEY AUTOINCREMENT NOT NULL,
  `name` text
);
--> statement-breakpoint
CREATE TABLE `book_to_author` (
  `author_id` integer,
  `book_id` integer,
  PRIMARY KEY(`book_id`, `author_id`)
);

Usage

export default {
  breakpoints: true
} satisfies Config;

tablesFilters

Type: string | string[]
Commands: push

tablesFilter param lets you filter tables with glob(opens in a new tab) syntax for db push command. It’s useful when you have only one database avaialable for several separate projects with separate sql schemas.

How to define multi-project tables with Drizzle ORM — see here.

Usage

export default {
  tablesFilter: ["project1_*"],
} satisfies Config;

In this case, only tables that start with project1_ will be synced with the database

schemaFilter

Type: string | string[]
Default: ["public"]
Commands: push, introspect
Available databases: PostgreSQL

The schemaFilter parameter allows you to define which schema in PostgreSQL should be used for either introspect or push commands. This parameter accepts a single schema as a string or an array of schemas as strings. No glob pattern is supported here. By default, drizzle will use the public schema for both commands, but you can add any schema you need.

For example, having schemaFilter: ["my_schema"] will only look for tables in both the database and drizzle schema that are a part of the my_schema schema.

Usage

export default {
  schemaFilter: ["public", "my_schema"],
} satisfies Config;

verbose

Type: boolean
Default: false
Commands: push

This command is used for drizzle-kit push commands and prints all statements that will be executed.

💡

Note: This command will only print the statements that should be executed. To approve them before applying, please refer to the strict command.

Usage

export default {
  verbose: true
} satisfies Config;

strict

Type: boolean
Default: false
Commands: push

This command is used for drizzle-kit push commands and will always ask for your confirmation, either to execute all statements needed to sync your schema with the database or not.

Usage

export default {
  strict: true
} satisfies Config;

introspect

Type: object

This section is a reference to introspect object inside drizzle.config.* file.

casing

Type: 'preserve' | 'camel'
Default: camel
Commands:

This command is responsible for a strategy for choosing column, table, constraint JS key names while introspecting your databse.

Usage

export default {
  introspect: {
    casing: 'preserve'
  }
} satisfies Config;

Example

preserve

camel

export const user_preferences = sqliteTable(
  "user_preferences",
  {
    id: integer("id").primaryKey(),
    name: integer("name").notNull(),
    user_id: integer("user_id").notNull(),
  }
);

export const userPreferences = sqliteTable(
  "user_preferences",
  {
    id: integer("id").primaryKey(),
    name: integer("name").notNull(),
    userId: integer("user_id").notNull(),
  }
);