Transactions in NestJS

Introduction

Transactions help keep your data safe by making sure a group of actions all finish together or none at all.

When saving multiple related records that must all succeed or fail together.

When updating several tables and you want to avoid partial changes.

When transferring money between accounts to prevent losing or duplicating funds.

When deleting data that affects other connected data and you want to keep consistency.

Syntax

NestJS

await connection.transaction(async (manager) => {
  // your database operations here
});

Use the transaction method from your database connection to run multiple operations safely.

Inside the transaction callback, use the provided manager to perform database actions.

Examples

This example saves a user and their profile together. If one save fails, none are saved.

NestJS

await connection.transaction(async (manager) => {
  await manager.save(user);
  await manager.save(profile);
});

This example updates an account balance inside a transaction to keep data safe.

NestJS

await connection.transaction(async (manager) => {
  const account = await manager.findOne(Account, { id: 1 });
  account.balance -= 100;
  await manager.save(account);
});

Sample Program

This service method creates a user and their profile inside a transaction. If saving either fails, no data is saved.

NestJS

import { Injectable } from '@nestjs/common';
import { DataSource } from 'typeorm';
import { User } from './user.entity';

@Injectable()
export class UserService {
  constructor(private dataSource: DataSource) {}

  async createUserWithProfile(userData, profileData) {
    return await this.dataSource.transaction(async (manager) => {
      const user = manager.create(User, userData);
      await manager.save(user);

      const profile = manager.create('Profile', { ...profileData, user });
      await manager.save(profile);

      return user;
    });
  }
}

Important Notes

Always use the transaction manager for all database operations inside the transaction callback.

Transactions help prevent data errors when multiple related changes happen together.

Summary

Transactions group multiple database actions to succeed or fail as one.

Use the transaction method from your DataSource in NestJS with an async callback.

Inside the callback, use the provided manager to run your database commands.

Practice

1. What is the main purpose of using transactions in a NestJS application with a database?

easy

A. To automatically generate database schemas

B. To ensure multiple database operations succeed or fail together

C. To speed up database queries by caching results

D. To log all database queries for debugging

Solution

1. Step 1: Understand what transactions do

   Transactions group multiple database actions so they either all succeed or all fail, preventing partial updates.

2. Step 2: Identify the purpose in NestJS context

   In NestJS, transactions keep data consistent by ensuring grouped operations are atomic.

3. Final Answer:

   To ensure multiple database operations succeed or fail together -> Option B

4. Quick Check:

   Transactions = atomic operations [OK]

Hint: Transactions keep grouped DB actions all-or-nothing [OK]

Common Mistakes:

• Thinking transactions speed up queries
• Confusing transactions with schema generation
• Assuming transactions are for logging only

2. Which of the following is the correct way to start a transaction using TypeORM's DataSource in NestJS?

easy

A. dataSource.transaction(async (manager) => { /* operations */ })

B. dataSource.startTransaction(() => { /* operations */ })

C. dataSource.begin(async () => { /* operations */ })

D. dataSource.runTransaction(() => { /* operations */ })

Solution

1. Step 1: Recall TypeORM transaction method

   TypeORM uses dataSource.transaction with an async callback to run transactional code.

2. Step 2: Check method names

   Only dataSource.transaction(async (manager) => { /* operations */ }) uses the correct method name and async callback pattern.

3. Final Answer:

   dataSource.transaction(async (manager) => { /* operations */ }) -> Option A

4. Quick Check:

   Correct method is transaction with async callback [OK]

Hint: Use dataSource.transaction with async callback [OK]

Common Mistakes:

• Using non-existent methods like startTransaction
• Omitting async keyword in callback
• Using wrong method names like begin or runTransaction

3. Given the following code snippet in a NestJS service using TypeORM, what will be the output if the second operation fails?

await dataSource.transaction(async (manager) => {
  await manager.save(user1);
  await manager.save(user2); // throws error
});
console.log('Transaction complete');

medium

A. Neither user1 nor user2 are saved, and an error is thrown before logging

B. Only user1 is saved, then 'Transaction complete' is logged

C. Both user1 and user2 are saved, then 'Transaction complete' is logged

D. 'Transaction complete' is logged before any save operations

Solution

1. Step 1: Understand transaction rollback behavior

   If any operation inside the transaction fails, all changes are rolled back, so no data is saved.

2. Step 2: Analyze the code flow

   The second save throws an error, so the transaction aborts and neither user1 nor user2 are saved. The error prevents reaching the console.log.

3. Final Answer:

   Neither user1 nor user2 are saved, and an error is thrown before logging -> Option A

4. Quick Check:

   Transaction rollback on error = no saves, error thrown [OK]

Hint: Error inside transaction rolls back all changes [OK]

Common Mistakes:

• Assuming partial saves happen before error
• Thinking console.log runs despite error
• Believing transaction commits despite failure

4. Identify the error in this NestJS transaction code snippet:

await dataSource.transaction(async (manager) => {
  manager.save(user1);
  await manager.save(user2);
});

medium

A. dataSource.transaction requires a synchronous callback

B. Transaction callback should not be async

C. Missing await before manager.save(user1)

D. manager.save cannot be used inside transactions

Solution

1. Step 1: Check async calls inside transaction

   All asynchronous operations inside the transaction callback must be awaited to ensure proper execution order.

2. Step 2: Identify missing await

   The first save call lacks await, so it may not complete before the second save starts, causing unexpected behavior.

3. Final Answer:

   Missing await before manager.save(user1) -> Option C

4. Quick Check:

   Await all async DB calls inside transaction [OK]

Hint: Always await async DB calls inside transaction [OK]

Common Mistakes:

• Omitting await on async calls
• Thinking transaction callback can't be async
• Believing manager.save is disallowed in transactions

5. You want to update two related entities atomically in NestJS using TypeORM transactions. Which approach correctly ensures both updates succeed or fail together?

async updateEntities(entityA, entityB) {
  await this.dataSource.transaction(async (manager) => {
    await manager.save(entityA);
    await manager.save(entityB);
  });
}

What is the best practice to handle errors and maintain data consistency?

hard

A. Manually call rollback() inside the transaction callback on error

B. Save entities without transactions and rely on database constraints

C. Use separate transactions for each save to isolate failures

D. Wrap the transaction call in try-catch and handle rollback automatically

Solution

1. Step 1: Understand transaction error handling in TypeORM

   TypeORM automatically rolls back if an error is thrown inside the transaction callback.

2. Step 2: Use try-catch outside transaction to handle errors

   Wrapping the transaction call in try-catch lets you catch errors and respond without manual rollback calls.

3. Final Answer:

   Wrap the transaction call in try-catch and handle rollback automatically -> Option D

4. Quick Check:

   Try-catch outside transaction handles errors; rollback is automatic [OK]

Hint: Use try-catch around transaction; rollback is automatic [OK]

Common Mistakes:

• Trying to manually rollback inside transaction callback
• Using separate transactions for related updates
• Ignoring transactions and relying only on DB constraints