TypeORM Setup for Your Web Application: Entities, Migrations, Repositories

When developing with Node.js and TypeScript, many face the same pain: manually writing SQL queries and synchronizing the schema with code. One mistake in an SQL query — and production goes down. TypeORM solves this problem, but its configuration requires accounting for a dozen nuances: from connection settings to migration generation. We will help you set up TypeORM for your project — from a simple server to a complex NestJS application. Our engineers have over 10 years of experience working with relational databases and guarantee the stability of your schema.

Why Standard TypeORM Setup Often Leads to Problems?

N+1 query — a classic problem when using ORM. Without optimization, loading 100 articles with authors will generate 101 queries. TypeORM makes it easy to set up eager loading via relations or use Query Builder for precise control. Another pain is synchronize: true in production: this leads to data loss when the schema changes. Proper migration configuration with diff file generation is the only safe way to update the database. The third issue is performance: unoptimized queries create extra load. TypeORM provides tools for monitoring and fine-tuning: subscribers for caching, connection pooling, slow query logging. Under a typical load of 10,000 requests per minute, a proper configuration reduces query count by 60%.

How We Configure TypeORM: A Case Study

From our practice: on a project with PostgreSQL and a load of 10,000 requests per minute, we faced performance degradation due to N+1 when loading related tags. Solution: we used Query Builder with leftJoinAndSelect and pagination via skip/take. Additionally, we configured a connection pool with max: 20 and idleTimeoutMillis: 30000, which reduced response time by 40%. Using Query Builder allowed executing a complex query 3 times faster compared to loading via relations. Proper connection pooling saved the client about 150,000 rubles per year on infrastructure. Order TypeORM setup — get a consultation from an engineer within an hour.

// db/data-source.ts import 'reflect-metadata' import { DataSource } from 'typeorm' import { User } from './entities/User' import { Post } from './entities/Post' export const AppDataSource = new DataSource({ type: 'postgres', url: process.env.DATABASE_URL, entities: [User, Post], migrations: ['dist/db/migrations/*.js'], migrationsTableName: 'migrations', synchronize: false, // NEVER true in production logging: process.env.NODE_ENV === 'development' ? ['query', 'error'] : ['error'], ssl: process.env.NODE_ENV === 'production' ? { rejectUnauthorized: false } : false, extra: { max: 20, idleTimeoutMillis: 30000, } }) // Initialization await AppDataSource.initialize() 

What Is Included in TypeORM Setup?

The work process consists of five stages:

1. Analysis — we examine the current database schema, load, and query profile.
2. Design — we choose a pattern (Active Record or Data Mapper), design entities and relationships.
3. Implementation — we configure the connection, write entities with decorators, create migrations and repositories.
4. Testing — we check performance, absence of N+1, and correctness of migrations.
5. Deployment and monitoring — we deploy to production, configure logging and alerts.

What is included in the result:

• DataSource configuration with pool and SSL (supports any database: PostgreSQL, MySQL, SQLite).
• Repositories and entities with indexes and relationships.
• Migrations (generation, running, rollback).
• Query Builder for typical queries.
• NestJS integration (module, services, controllers).
• Subscribers for caching and notifications.
• API and architecture documentation.
• 30-day warranty after delivery — we fix bugs for free.

Pattern Comparison and Entity Selection

Our engineers always recommend Data Mapper for projects with business logic, as it gives more control and simplifies testing. But if you need rapid prototyping — Active Record is also a working option. Proper entity selection is critical for performance. Use @PrimaryGeneratedColumn('uuid') for distributed systems, @Index() for frequently filtered fields. A repository is a layer between business logic and the database. Wrap all queries to an entity in a repository: this makes the code reusable and testable.

Example repository with pagination:

const postRepository = AppDataSource.getRepository(Post) async function findPosts(opts: { page: number; limit: number; search?: string }) { const { page, limit, search } = opts const qb = postRepository.createQueryBuilder('post') .leftJoinAndSelect('post.author', 'author') .leftJoinAndSelect('post.tags', 'tag') .where('post.published = :published', { published: true }) .orderBy('post.createdAt', 'DESC') .skip((page - 1) * limit) .take(limit) if (search) { qb.andWhere( 'post.title ILIKE :search OR post.content ILIKE :search', { search: `%${search}%` } ) } const [items, total] = await qb.getManyAndCount() return { items, total, pages: Math.ceil(total / limit) } } // Complex aggregates const stats = await AppDataSource.query(` SELECT date_trunc('week', created_at) AS week, count(*) AS posts, count(*) FILTER (WHERE published = true) AS published FROM posts WHERE created_at >= now() - interval '90 days' GROUP BY 1 ORDER BY 1 `) 

Why Migrations Are Important for Production?

According to official TypeORM documentation, migrations are the only safe way to manage the schema in production. Migations allow versioning database schema changes and applying them sequentially. TypeORM can automatically generate migrations based on changes in entities. This saves up to 50% of time on developing migrations compared to writing SQL manually. Example of generation and application:

# Generate migration from schema diff npx typeorm migration:generate -n AddUserProfile -d dist/db/data-source.js # Create an empty migration manually npx typeorm migration:create -n AddIndexes # Apply npx typeorm migration:run -d dist/db/data-source.js # Revert the last one npx typeorm migration:revert -d dist/db/data-source.js 

// db/migrations/1234567890-AddUserProfile.ts import { MigrationInterface, QueryRunner } from 'typeorm' export class AddUserProfile1234567890 implements MigrationInterface { public async up(queryRunner: QueryRunner): Promise<void> { await queryRunner.query(` CREATE TABLE profiles ( id UUID PRIMARY KEY DEFAULT gen_random_uuid(), user_id UUID NOT NULL UNIQUE REFERENCES users(id) ON DELETE CASCADE, bio TEXT, avatar_url VARCHAR(500), updated_at TIMESTAMPTZ NOT NULL DEFAULT now() ) `) } public async down(queryRunner: QueryRunner): Promise<void> { await queryRunner.query(`DROP TABLE profiles`) } } 

NestJS Integration and Event Subscribers

In NestJS, setting up TypeORM boils down to importing TypeOrmModule with configuration. Additionally, you can inject repositories via the @InjectRepository decorator. Subscribers allow reacting to entity changes: after insert, update search index; after publication, send a notification. This is a powerful tool if not overused. Example of NestJS setup:

// app.module.ts import { TypeOrmModule } from '@nestjs/typeorm' @Module({ imports: [ TypeOrmModule.forRootAsync({ inject: [ConfigService], useFactory: (config: ConfigService) => ({ type: 'postgres', url: config.get('DATABASE_URL'), entities: [__dirname + '/**/*.entity{.ts,.js}'], migrations: [__dirname + '/db/migrations/*{.ts,.js}'], migrationsRun: true, synchronize: false, }) }), TypeOrmModule.forFeature([User, Post]) ] }) export class AppModule {} 

Estimated Timeframes and Cost

Basic TypeORM setup with entities, migrations, and repositories: 1–2 days. Integration into a NestJS project with modules and tests: 2–3 days. Migrating an existing project from another ORM to TypeORM: 3–5 days. The cost of basic TypeORM configuration depends on project complexity. Contact us for a free estimate — it takes no more than an hour.