drizzle-orm-sqlite

name	drizzle-orm-sqlite
description	Use Drizzle ORM with SQLite and node:sqlite; load when defining SQLite schemas, building typed queries, using transactions, or testing Drizzle persistence without Drizzle Kit.

Use the installed Drizzle ORM API without assuming a Drizzle Kit workflow.

Process

1. Identify the installed boundary.

   • Read the package manifest and lockfile for the exact drizzle-orm version.
   • Identify the driver, connection owner, schema exports, and sync or async API style.
   • Prefer installed types when release-candidate APIs differ from current docs.
   • Completion: the ORM version, SQLite driver, and connection lifecycle are known.

2. Define the schema.

   • Use sqliteTable and SQLite column builders from drizzle-orm/sqlite-core.
   • Express keys, constraints, defaults, checks, and indexes explicitly.
   • Align TypeScript declarations with the schema applied to the database.
   • Treat declarations as metadata; they do not alter an existing database alone.
   • Completion: declarations match the database objects and constraints the code expects.

3. Build typed data access.

   • Prefer Drizzle select, insert, update, delete, and relational query builders.
   • Import operators such as eq, and, or, inArray, and desc from drizzle-orm.
   • Use sql fragments for dialect behavior the typed builders cannot express directly.
   • Bind external values through Drizzle instead of interpolating SQL strings.
   • Select only the columns needed when rows contain large or sensitive values.
   • Completion: inputs, results, ordering, and empty-result behavior are explicit.

4. Preserve transaction semantics.

   • Use db.transaction for writes that must commit or roll back together.
   • Keep all correlated reads and writes on the transaction handle.
   • Match synchronous node:sqlite behavior to the installed Drizzle API.
   • Make retry and idempotency decisions at the application boundary.
   • Completion: partial failure cannot leave the affected invariant half-applied.

5. Handle schema evolution deliberately.

   • Inspect installed tooling before choosing migration, push, or initialization SQL.
   • Keep Drizzle Kit commands outside the workflow when drizzle-kit is not installed.
   • Pair persisted schema changes with a compatible upgrade or rebuild strategy.
   • Completion: both new and existing databases have a defined schema path.

6. Verify through SQLite.

   • Exercise queries against a temporary real SQLite database.
   • Test constraints, transaction rollback, ordering, and restart behavior when relevant.
   • Close database handles and remove temporary state during cleanup.
   • Inspect sqlite_master or pragmas when database structure is part of the behavior.
   • Completion: tests observe behavior across the actual Drizzle and SQLite boundary.

Official references

• Node SQLite driver
• SQLite schema declarations
• Select queries
• Transactions