Skip to content

database.md

Latest commit

 

History

History
451 lines (336 loc) · 15.8 KB

database.md

File metadata and controls

451 lines (336 loc) · 15.8 KB
Raw
outline deep
<script setup> const dbFileItems = [ { name: "db/", children: [ { name: "schema.ts", description: "Table definitions (source of truth)" }, { name: "schema/", description: "Split schema files (optional)", children: [ { name: "users.ts" }, { name: "posts.ts" }, ], }, { name: "migrations/", description: "Generated SQL migration files", children: [ { name: "20260410161500_create_users.sql" }, { name: "20260410161501_add_posts.sql" }, ], }, { name: "seed.ts", description: "Programmatic seed script (optional)" }, { name: "seed.sql", description: "Raw SQL seed file (optional alternative)" }, ], }, ] </script>

Database

Define your tables in TypeScript, then query them with db from void/db. Void includes Drizzle ORM and configures the connection, so you can use the same schema and query workflow with D1, PostgreSQL, or MySQL.

Choosing a Dialect

Choose a database during void init, or set database in void.config.ts:

D1 (SQLite) PostgreSQL MySQL
Config Default "database": "pg" "database": "mysql"
Managed by Void Bring your own database Bring your own database
Connection Automatic D1 binding Hyperdrive Hyperdrive

A Void platform keeps Hyperdrive query caching off for your database connection, so reads reflect completed writes. An existing connection is updated on the project's next deploy.

Schema Definition

Define your tables in db/schema.ts using column helpers from void/schema-d1, void/schema-pg, or void/schema-mysql.

::: code-group

// db/schema.ts
import { sql } from 'void/db';
import { sqliteTable, text, integer } from 'void/schema-d1';

export const users = sqliteTable('users', {
  id: integer('id').primaryKey({ autoIncrement: true }),
  name: text('name').notNull(),
  email: text('email').notNull().unique(),
  role: text('role').notNull().default('user'),
  createdAt: text('created_at')
    .notNull()
    .default(sql`(datetime('now'))`),
});
// db/schema.ts
import { pgTable, serial, text, timestamp } from 'void/schema-pg';

export const users = pgTable('users', {
  id: serial('id').primaryKey(),
  name: text('name').notNull(),
  email: text('email').notNull().unique(),
  role: text('role').notNull().default('user'),
  createdAt: timestamp('created_at').notNull().defaultNow(),
});
// db/schema.ts
import { int, mysqlTable, text, timestamp } from 'void/schema-mysql';

export const users = mysqlTable('users', {
  id: int('id').autoincrement().primaryKey(),
  name: text('name').notNull(),
  email: text('email').notNull(),
  createdAt: timestamp('created_at').notNull().defaultNow(),
});

:::

The schema helpers are included in void. You can write tables yourself or generate a starting point with void gen model.

You can also split your schema across multiple files under db/schema/ and re-export from a barrel file:

// db/schema/users.ts
export const users = ...;

// db/schema/posts.ts
export const posts = ...;

// db/schema.ts (barrel)
export * from "./schema/users";
export * from "./schema/posts";

Column casing

By default, a column declared without a name, such as createdAt: timestamp(), uses its key as the column name. Set database.casing to 'snake_case' to store these columns as created_at instead. It applies to void/db queries and to void db commands such as generate, push, and seed. Columns with an explicit name, such as text('created_at'), keep that name.

// void.config.ts
import { defineConfig } from 'void/config';

export default defineConfig({
  database: { dialect: 'pg', casing: 'snake_case' },
});
// db/schema.ts
import { pgTable, serial, text, timestamp } from 'void/schema-pg';

export const posts = pgTable('posts', {
  id: serial().primaryKey(),
  title: text().notNull(),
  createdAt: timestamp().notNull().defaultNow(), // created_at
});

Omit dialect for D1. If your database already has tables, run void db generate after you change casing and review the migration before you apply it.

Querying

Void re-exports all common Drizzle query operators from void/db so you don't need to install drizzle-orm separately. Import db from void/db and your tables from @schema:

import { users } from '@schema';
import { db } from 'void/db';

::: info What is @schema? @schema is a Vite path alias that the Void plugin configures automatically. It points to your db/schema.ts file, or db/schema/ if you split tables across files. You can just use it to import table definitions. :::

List rows

const allUsers = await db.select().from(users);

Filter with where

import { users } from '@schema';
import { db, eq, and, or } from 'void/db';

// Single condition
const user = await db.select().from(users).where(eq(users.id, 1));

// Multiple conditions
const admins = await db
  .select()
  .from(users)
  .where(and(eq(users.role, 'admin'), eq(users.name, 'Alice')));

// OR conditions
const result = await db
  .select()
  .from(users)
  .where(or(eq(users.role, 'admin'), eq(users.role, 'editor')));

Insert

// Single row
await db.insert(users).values({
  name: 'Alice',
  email: 'alice@example.com',
});

// Insert and return the created row
const [created] = await db
  .insert(users)
  .values({ name: 'Bob', email: 'bob@example.com' })
  .returning();

// Multiple rows
await db.insert(users).values([
  { name: 'Alice', email: 'alice@example.com' },
  { name: 'Bob', email: 'bob@example.com' },
]);

Update

await db.update(users).set({ role: 'admin' }).where(eq(users.id, 1));

// Update and return the modified row
const [updated] = await db.update(users).set({ role: 'admin' }).where(eq(users.id, 1)).returning();

Delete

await db.delete(users).where(eq(users.id, 1));

Column selection

const names = await db.select({ name: users.name, email: users.email }).from(users);
// names: { name: string; email: string }[]

Ordering and pagination

import { users } from '@schema';
import { db, desc } from 'void/db';

const page = await db.select().from(users).orderBy(desc(users.createdAt)).limit(10).offset(20);

Joins

import { users, posts } from '@schema';
import { db, eq } from 'void/db';

const results = await db
  .select({
    id: posts.id,
    title: posts.title,
    author: users.name,
  })
  .from(posts)
  .innerJoin(users, eq(posts.userId, users.id));

Relational queries

If your schema defines relations, you can use Drizzle's relational query API:

const usersWithPosts = await db.query.users.findMany({
  with: { posts: true },
});

Seeding

Use void db seed to reset your local database, re-apply migrations, and then run a seed file.

Use db/seed.ts for a programmatic seed or db/seed.sql for SQL. JavaScript and .mts / .mjs files also work. If more than one seed file exists, choose one with --file <path>.

Programmatic seeding

Use db/seed.ts when you want to generate seed data in code. Export a default function or a named seed function.

// db/seed.ts
import { defineSeed } from 'void/seed';

export default defineSeed<typeof import('./schema')>(async ({ db, schema }) => {
  const rows = Array.from({ length: 100 }, (_, i) => ({
    text: `Seed message ${i + 1}`,
  }));

  await db.insert(schema.messages).values(rows);
});

The seed context includes:

  • dialect: "sqlite", "postgresql", or "mysql"
  • db: a Drizzle instance for the local database
  • schema: the exports from your db/schema.ts or db/schema/ modules

SQL seeding

If you prefer raw SQL, keep using db/seed.sql:

INSERT INTO messages (text) VALUES ('Hello from SQL');

Schema-Derived Validators

Derive request validators from your tables with void/drizzle-zod, void/drizzle-valibot, or void/drizzle-arktype. Add the validator after your table definition:

::: code-group

// db/schema.ts
import { createInsertSchema } from 'void/drizzle-zod';

export const insertUserSchema = createInsertSchema(users, {
  name: (schema) => schema.min(1),
  email: (schema) => schema.email(),
});
// db/schema.ts
import { pipe, minLength, email } from 'valibot';
import { createInsertSchema } from 'void/drizzle-valibot';

export const insertUserSchema = createInsertSchema(users, {
  name: (schema) => pipe(schema, minLength(1)),
  email: (schema) => pipe(schema, email()),
});
// db/schema.ts
import { type } from 'arktype';
import { createInsertSchema } from 'void/drizzle-arktype';

export const insertUserSchema = createInsertSchema(users, {
  name: type('string > 0'),
  email: type('string.email'),
});

:::

createInsertSchema generates a validator that matches $inferInsert. Columns with defaults or auto-increments become optional, while NOT NULL columns stay required. The optional second argument lets you refine individual columns, either with a callback that receives the generated schema or by passing a type directly to override a field.

Use the derived schema in your route handlers with withValidator():

// routes/api/users/index.ts
import { users, insertUserSchema } from '@schema';
import { defineHandler } from 'void';
import { db } from 'void/db';

export const POST = defineHandler.withValidator({
  body: insertUserSchema,
})(async (c, { body }) => {
  const [created] = await db.insert(users).values(body).returning();
  return created;
});

Three functions are available:

Function Purpose
createInsertSchema Validates insert data, excluding auto-generated columns and applying defaults
createSelectSchema Matches the shape of selected rows
createUpdateSchema Like insert but all fields are optional (partial update)

Void already bundles the Drizzle adapters. Install the validator library itself if you use its direct APIs in your schema file:

npm install zod
# or
npm install valibot
# or
npm install arktype

CLI Commands

Command Purpose
void db push Apply schema changes locally without migration files
void db generate Generate SQL migrations
void db migrate Apply pending migrations locally
void db status Check schema drift and pending migrations
void db seed Reset the local database and run a seed file

See the CLI reference for connection setup, SQL execution, Studio, exports, and other commands.

Adopt an existing Drizzle PostgreSQL database {#adopt-existing-drizzle}

You can keep a PostgreSQL database that Drizzle has already migrated. Void tracks applied files separately, so record the existing history before the first Void deployment.

  1. Back up the database and pause deployments. Try the procedure on a restored copy first.
  2. Set database: 'pg' in void.config.ts. Keep your existing schema, SQL filenames, and db/migrations/meta/ directory. Do not regenerate or rename migrations during adoption.
  3. Check the database's Drizzle history against meta/_journal.json and your SQL files. Match each applied row's created_at to the journal entry's when, and its hash to the SHA-256 of the corresponding SQL file. Drizzle normally stores this history in drizzle.__drizzle_migrations; use your configured schema and table if different. If a file cannot be verified, resolve the discrepancy before continuing.
SELECT hash, created_at FROM drizzle.__drizzle_migrations ORDER BY created_at;
  1. Save the following as adopt-void.sql, replacing the example filenames with only the migrations verified as already applied to this database, including .sql. Leave new, unapplied files out of the list.
BEGIN;
CREATE TABLE IF NOT EXISTS _void_migrations (
  name TEXT PRIMARY KEY,
  applied_at TEXT NOT NULL
);
INSERT INTO _void_migrations (name, applied_at)
VALUES
  ('0000_create_users.sql', CURRENT_TIMESTAMP::text),
  ('0001_add_email.sql', CURRENT_TIMESTAMP::text)
ON CONFLICT (name) DO NOTHING;
COMMIT;
  1. Run the file against the existing database with your PostgreSQL client, stopping on errors. Use the same database and PostgreSQL search path that Void will use. This records the verified filenames without executing their SQL or changing Drizzle's history.
  2. Configure Void's connection to that database and run void db status. Verify that only new migrations are pending, then deploy. For direct Cloudflare deployments, provide that database's DATABASE_URL in the shell or CI environment running void deploy.

Repeat the verification and adoption separately for each database. An entry in one environment's history does not prove that another environment has applied it. Leave Drizzle's history table intact, and use Void for subsequent migrations.

Hand-written migrations {#hand-written-migrations}

Deploy compares your Drizzle schema with the snapshots in db/migrations/meta/. If you have SQL migrations without this metadata, running void db generate would create another migration for the same tables.

To keep an existing migration history:

  1. Make your Drizzle schema match the tables your SQL migrations create.
  2. In a scratch copy of the project, remove db/migrations/ and run void db generate.
  3. Copy the generated meta/ folder into your original migration directory. In _journal.json, replace entries with an entry for each existing SQL file in order. Use consecutive idx values starting at 0, the file name without .sql as tag, increasing timestamps as when, and the generated entry's version. Set breakpoints to whether the file uses --> statement-breakpoint lines.
  4. Rename the generated snapshot to <prefix>_snapshot.json, where <prefix> is the part of the last migration tag before its first _. For 001_create_users.sql, use 001_snapshot.json.

Review and commit the metadata, then run void deploy again. If no database has applied the SQL yet, you can instead replace it with migrations from void db generate.

For D1 migrations containing triggers, and MySQL migrations containing triggers, procedures, functions or events, add a --> statement-breakpoint line after each complete SQL statement, including the trigger's closing END;. Keep the trigger body together.

Scaffolding

The void gen model command generates dialect-appropriate sqliteTable, pgTable, or mysqlTable definitions.

void gen model posts title:string body:text published:boolean

This creates:

  1. db/schema/posts.ts: a Drizzle table definition with id, createdAt, updatedAt, and your columns
  2. Updates db/schema.ts with export * from "./schema/posts"
  3. routes/api/posts/index.ts: GET for list and POST for insert with validation
  4. routes/api/posts/[id].ts: GET by id with 404 handling

See the CLI reference for the full list of generators.