prisma

Prisma Skill

You are working with Prisma ORM. Apply these conventions.

Schema (schema.prisma)

• Use @@map("snake_case_table") on models to keep DB tables in snake_case while Prisma models stay PascalCase.
• Use @map("snake_case_col") on fields for the same reason.
• Always add @@index([field]) for columns used in WHERE, ORDER BY, or as foreign keys.
• Use @db.Text for long strings; use @db.VarChar(255) for constrained strings on PostgreSQL.
• Use @updatedAt for auto-updating timestamps.
• Use enum for fixed value sets; map to DB enums or use a string with @db.VarChar.

Migrations

• Run npx prisma migrate dev --name descriptive-name during development — generates SQL and applies it.
• Run npx prisma migrate deploy in CI/production — never migrate dev in production.
• Never edit migration files by hand; instead, update schema and create a new migration.
• Use npx prisma db push for rapid prototyping when migration history doesn't matter.
• Introspect existing databases with npx prisma db pull.

Prisma Client

• Instantiate a single PrismaClient and reuse it (module singleton pattern):

  // lib/prisma.ts
  import { PrismaClient } from '@prisma/client';
  const globalForPrisma = globalThis as unknown as { prisma: PrismaClient };
  export const prisma = globalForPrisma.prisma ?? new PrismaClient();
  if (process.env.NODE_ENV !== 'production') globalForPrisma.prisma = prisma;
  

• Always call $disconnect() in CLI scripts and tests; managed services handle it automatically.
• Use $transaction([…]) for atomic operations or $transaction(async (tx) => { … }) for interactive transactions.

Queries

• Use select to fetch only needed fields — never over-fetch entire records.
• Use include for eager-loading relations; use nested select inside include to limit relation fields.
• Use where: { id: { in: ids } } for bulk lookups.
• Paginate with skip and take; use cursor-based pagination for large datasets.