Point migrations at a folder of .sql files. Each deploy runs the new ones, in order.
Apply a folder
const db = yield* Turso.Database("Db", {
group: group.name,
migrations: "./migrations",
}); -- migrations/0001_users.sql
CREATE TABLE users (id INTEGER PRIMARY KEY, name TEXT NOT NULL); If any statement in a file fails, none of that file's changes are kept and the deploy fails.
Alchemy keeps track of which files ran in the __alchemy_migrations table.
Make migrations with Drizzle
const schema = yield* Drizzle.Schema("Schema", {
schema: "./src/schema.ts",
out: "./migrations",
dialect: "sqlite",
});
const db = yield* Turso.Database("Db", {
group: group.name,
migrations: schema,
}); Drizzle.Schema writes a migration when src/schema.ts changes. Add Drizzle.providers() to your Stack and commit the ./migrations folder.
Rather run drizzle-kit generate yourself? Pass the folder instead: migrations: "./migrations".
Already used drizzle-kit on this database? Alchemy picks up its history on the first deploy.
Use a different tracking table
migrations: { dir: "./migrations", table: "schema_migrations" }, Seed with SQL files
const db = yield* Turso.Database("Db", {
group: group.name,
migrations: "./migrations",
importFiles: ["./seed/users.sql"],
}); importFiles run after migrations. A file runs again only if you change it.
Don't edit old migrations
Changing or deleting a file that already ran makes the deploy fail. Add a new file instead.
Where next
- SQL migrations — how migrations work for every database
- Database per tenant — updating many databases