Skip to main content
Introducing packages.sweber.dev
Documentation menuDrizzle and Prisma

Drizzle and Prisma

Use Logarithm next to Drizzle ORM or Prisma, with your existing database connection.

Logarithm talks to the database through a small client interface, not through an ORM, so it works next to any ORM. The audit table is managed by Logarithm, and your ORM keeps managing the rest.

Drizzle ORM

Drizzle's node-postgres and postgres-js drivers wrap a pool. Use the same pool for Logarithm:

import { drizzle } from "drizzle-orm/node-postgres"
import { Pool } from "pg"
import { createAuditLog } from "@sweberdev/logarithm"
import { migratePostgres, postgresStore } from "@sweberdev/logarithm/postgres"

const pool = new Pool({ connectionString: process.env.DATABASE_URL })
export const db = drizzle(pool)

await migratePostgres({ client: pool })
export const audit = createAuditLog({ store: postgresStore({ client: pool }) })

To keep the table in Drizzle's migration history, generate an empty migration with drizzle-kit generate --custom --name audit_log and paste the output of postgresSchema() into the file. Drizzle then applies it with your other migrations, and you do not need to call migratePostgres.

Prisma

Prisma Client does not expose a raw query(sql, values) method, so give Logarithm the underlying pg pool (Prisma needs a connection string anyway) or use Prisma's driver adapter:

import { Pool } from "pg"
import { PrismaPg } from "@prisma/adapter-pg"
import { PrismaClient } from "@prisma/client"
import { createAuditLog } from "@sweberdev/logarithm"
import { postgresStore } from "@sweberdev/logarithm/postgres"

const pool = new Pool({ connectionString: process.env.DATABASE_URL })
export const prisma = new PrismaClient({ adapter: new PrismaPg(pool) })
export const audit = createAuditLog({ store: postgresStore({ client: pool }) })

To keep the table in Prisma's migrations, run prisma migrate dev --create-only --name audit_log, paste postgresSchema() into the generated migration.sql, then prisma migrate dev. Do not model the audit table in schema.prisma; Prisma would try to manage it.

Recording the change and the audit entry together

record() takes before and after objects, so the usual pattern is to read the row, update it and record the diff:

const before = await prisma.project.findUniqueOrThrow({ where: { id } })
const after = await prisma.project.update({ where: { id }, data })
await audit.record({
  action: "project.updated",
  actor: { id: session.userId },
  targets: [{ type: "project", id, name: after.name }],
  before,
  after,
})