Skip to content

Migrations

An .sq file always describes how to create the latest schema in an empty database. If your database is currently on an earlier version, migration files bring those databases up-to-date. Migration files are stored in the same sqldelight folder as your .sq files:

src
└─ main
   └─ sqdelight
      ├─ com/example/hockey
      |  ├─ Team.sq
      |  └─ Player.sq
      └─ migrations
         ├─ 1.sqm
         └─ 2.sqm

If the driver supports it, migrations are run in a transaction. You should not surround your migrations in BEGIN/END TRANSACTION, as this can cause a crash with some drivers.

Versioning

The first version of the schema is 1. Migration files are named <version to upgrade from>.sqm. To migrate to version 2, put migration statements in 1.sqm:

ALTER TABLE hockeyPlayer ADD COLUMN draft_year INTEGER;
ALTER TABLE hockeyPlayer ADD COLUMN draft_order INTEGER;

These SQL statements are run by the Database.Schema.migrate() method. Migration files go in the same source set as your .sq files.

Verifying Migrations

You can also place a .db file in the src/main/sqldelight folder of the same <version number>.db format. If there is a .db file present, a new verifySqlDelightMigration task will be added to the gradle project, and it will run as part of the check task, meaning your migrations will be verified against that .db file. It confirms that the migrations yield a database with the latest schema.

To generate a .db file from your latest schema, run the generateSqlDelightSchema task, which is available once you specify a schemaOutputDirectory, as described in the gradle.md. You should probably do this before you create your first migration.

Code Migrations

If you run your migration from code and would like to perform data migrations you can use the Database.Schema.migrate api:

Database.Schema.migrate(
    driver = database,
    oldVersion = 0,
    newVersion = Database.Schema.version,
    AfterVersion(3) { driver -> driver.execute(null, "INSERT INTO test (value) VALUES('hello')", 0) },
)

In the following example, if you have 1.sqm, 2.sqm, 3.sqm, 4.sqm, and 5.sqm as migrations, the above callback will happen after 3.sqm completes when the database is on version 4. After the callback it will resume at 4.sqm and complete the remaining migrations, in this case 4.sqm and 5.sqm, meaning the final database version is 6.

If you are using an JdbcSqliteDriver you can pass the schema and callbacks in during the driver's creation. It uses PRAGMA user_version to store current version of schema in database.

val driver: SqlDriver = JdbcSqliteDriver(
    url = "jdbc:sqlite:test.db",
    properties = Properties(),
    schema = Database.Schema,
    callbacks = arrayOf(
        AfterVersion(3) { driver -> driver.execute(null, "INSERT INTO test (value) VALUES('hello')", 0) }
    )
)