- 1. API with NestJS #1. Controllers, routing and the module structure
- 2. API with NestJS #2. Setting up a PostgreSQL database with TypeORM
- 3. API with NestJS #3. Authenticating users with bcrypt, Passport, JWT, and cookies
- 4. API with NestJS #4. Error handling and data validation
- 5. API with NestJS #5. Serializing the response with interceptors
- 6. API with NestJS #6. Looking into dependency injection and modules
- 7. API with NestJS #7. Creating relationships with Postgres and TypeORM
- 8. API with NestJS #8. Writing unit tests
- 9. API with NestJS #9. Testing services and controllers with integration tests
- 10. API with NestJS #10. Uploading public files to Amazon S3
- 11. API with NestJS #11. Managing private files with Amazon S3
- 12. API with NestJS #12. Introduction to Elasticsearch
- 13. API with NestJS #13. Implementing refresh tokens using JWT
- 14. API with NestJS #14. Improving performance of our Postgres database with indexes
- 15. API with NestJS #15. Defining transactions with PostgreSQL and TypeORM
- 16. API with NestJS #16. Using the array data type with PostgreSQL and TypeORM
- 17. API with NestJS #17. Offset and keyset pagination with PostgreSQL and TypeORM
- 18. API with NestJS #18. Exploring the idea of microservices
- 19. API with NestJS #19. Using RabbitMQ to communicate with microservices
- 20. API with NestJS #20. Communicating with microservices using the gRPC framework
- 21. API with NestJS #21. An introduction to CQRS
- 22. API with NestJS #22. Storing JSON with PostgreSQL and TypeORM
- 23. API with NestJS #23. Implementing in-memory cache to increase the performance
- 24. API with NestJS #24. Cache with Redis. Running the app in a Node.js cluster
- 25. API with NestJS #25. Sending scheduled emails with cron and Nodemailer
- 26. API with NestJS #26. Real-time chat with WebSockets
- 27. API with NestJS #27. Introduction to GraphQL. Queries, mutations, and authentication
- 28. API with NestJS #28. Dealing in the N + 1 problem in GraphQL
- 29. API with NestJS #29. Real-time updates with GraphQL subscriptions
- 30. API with NestJS #30. Scalar types in GraphQL
- 31. API with NestJS #31. Two-factor authentication
- 32. API with NestJS #32. Introduction to Prisma with PostgreSQL
- 33. API with NestJS #33. Managing PostgreSQL relationships with Prisma
- 34. API with NestJS #34. Handling CPU-intensive tasks with queues
- 35. API with NestJS #35. Using server-side sessions instead of JSON Web Tokens
- 36. API with NestJS #36. Introduction to Stripe with React
- 37. API with NestJS #37. Using Stripe to save credit cards for future use
- 38. API with NestJS #38. Setting up recurring payments via subscriptions with Stripe
- 39. API with NestJS #39. Reacting to Stripe events with webhooks
- 40. API with NestJS #40. Confirming the email address
- 41. API with NestJS #41. Verifying phone numbers and sending SMS messages with Twilio
- 42. API with NestJS #42. Authenticating users with Google
- 43. API with NestJS #43. Introduction to MongoDB
- 44. API with NestJS #44. Implementing relationships with MongoDB
- 45. API with NestJS #45. Virtual properties with MongoDB and Mongoose
- 46. API with NestJS #46. Managing transactions with MongoDB and Mongoose
- 47. API with NestJS #47. Implementing pagination with MongoDB and Mongoose
- 48. API with NestJS #48. Definining indexes with MongoDB and Mongoose
- 49. API with NestJS #49. Updating with PUT and PATCH with MongoDB and Mongoose
- 50. API with NestJS #50. Introduction to logging with the built-in logger and TypeORM
- 51. API with NestJS #51. Health checks with Terminus and Datadog
- 52. API with NestJS #52. Generating documentation with Compodoc and JSDoc
- 53. API with NestJS #53. Implementing soft deletes with PostgreSQL and TypeORM
- 54. API with NestJS #54. Storing files inside a PostgreSQL database
- 55. API with NestJS #55. Uploading files to the server
- 56. API with NestJS #56. Authorization with roles and claims
- 57. API with NestJS #57. Composing classes with the mixin pattern
- 58. API with NestJS #58. Using ETag to implement cache and save bandwidth
- 59. API with NestJS #59. Introduction to a monorepo with Lerna and Yarn workspaces
- 60. API with NestJS #60. The OpenAPI specification and Swagger
- 61. API with NestJS #61. Dealing with circular dependencies
- 62. API with NestJS #62. Introduction to MikroORM with PostgreSQL
- 63. API with NestJS #63. Relationships with PostgreSQL and MikroORM
- 64. API with NestJS #64. Transactions with PostgreSQL and MikroORM
- 65. API with NestJS #65. Implementing soft deletes using MikroORM and filters
- 66. API with NestJS #66. Improving PostgreSQL performance with indexes using MikroORM
- 67. API with NestJS #67. Migrating to TypeORM 0.3
- 68. API with NestJS #68. Interacting with the application through REPL
- 69. API with NestJS #69. Database migrations with TypeORM
- 70. API with NestJS #70. Defining dynamic modules
- 71. API with NestJS #71. Introduction to feature flags
- 72. API with NestJS #72. Working with PostgreSQL using raw SQL queries
- 73. API with NestJS #73. One-to-one relationships with raw SQL queries
- 74. API with NestJS #74. Designing many-to-one relationships using raw SQL queries
- 75. API with NestJS #75. Many-to-many relationships using raw SQL queries
- 76. API with NestJS #76. Working with transactions using raw SQL queries
- 77. API with NestJS #77. Offset and keyset pagination with raw SQL queries
- 78. API with NestJS #78. Generating statistics using aggregate functions in raw SQL
- 79. API with NestJS #79. Implementing searching with pattern matching and raw SQL
- 80. API with NestJS #80. Updating entities with PUT and PATCH using raw SQL queries
- 81. API with NestJS #81. Soft deletes with raw SQL queries
- 82. API with NestJS #82. Introduction to indexes with raw SQL queries
- 83. API with NestJS #83. Text search with tsvector and raw SQL
- 84. API with NestJS #84. Implementing filtering using subqueries with raw SQL
- 85. API with NestJS #85. Defining constraints with raw SQL
- 86. API with NestJS #86. Logging with the built-in logger when using raw SQL
- 87. API with NestJS #87. Writing unit tests in a project with raw SQL
- 88. API with NestJS #88. Testing a project with raw SQL using integration tests
- 89. API with NestJS #89. Replacing Express with Fastify
- 90. API with NestJS #90. Using various types of SQL joins
- 91. API with NestJS #91. Dockerizing a NestJS API with Docker Compose
- 92. API with NestJS #92. Increasing the developer experience with Docker Compose
- 93. API with NestJS #93. Deploying a NestJS app with Amazon ECS and RDS
- 94. API with NestJS #94. Deploying multiple instances on AWS with a load balancer
- 95. API with NestJS #95. CI/CD with Amazon ECS and GitHub Actions
- 96. API with NestJS #96. Running unit tests with CI/CD and GitHub Actions
- 97. API with NestJS #97. Introduction to managing logs with Amazon CloudWatch
- 98. API with NestJS #98. Health checks with Terminus and Amazon ECS
- 99. API with NestJS #99. Scaling the number of application instances with Amazon ECS
- 100. API with NestJS #100. The HTTPS protocol with Route 53 and AWS Certificate Manager
- 101. API with NestJS #101. Managing sensitive data using the AWS Secrets Manager
- 102. API with NestJS #102. Writing unit tests with Prisma
- 103. API with NestJS #103. Integration tests with Prisma
- 104. API with NestJS #104. Writing transactions with Prisma
- 105. API with NestJS #105. Implementing soft deletes with Prisma and middleware
- 106. API with NestJS #106. Improving performance through indexes with Prisma
- 107. API with NestJS #107. Offset and keyset pagination with Prisma
- 108. API with NestJS #108. Date and time with Prisma and PostgreSQL
- 109. API with NestJS #109. Arrays with PostgreSQL and Prisma
- 110. API with NestJS #110. Managing JSON data with PostgreSQL and Prisma
- 111. API with NestJS #111. Constraints with PostgreSQL and Prisma
- 112. API with NestJS #112. Serializing the response with Prisma
- 113. API with NestJS #113. Logging with Prisma
- 114. API with NestJS #114. Modifying data using PUT and PATCH methods with Prisma
- 115. API with NestJS #115. Database migrations with Prisma
- 116. API with NestJS #116. REST API versioning
- 117. API with NestJS #117. CORS – Cross-Origin Resource Sharing
- 118. API with NestJS #118. Uploading and streaming videos
- 119. API with NestJS #119. Type-safe SQL queries with Kysely and PostgreSQL
- 120. API with NestJS #120. One-to-one relationships with the Kysely query builder
- 121. API with NestJS #121. Many-to-one relationships with PostgreSQL and Kysely
In this series, we’ve implemented deleting functionalities for various entities. So far, it has always meant removing records permanently from our database. Instead, we can perform soft deletes. When we delete a record using a soft delete, we only mark it as deleted.
You can find the code from this series in this repository.
Soft deletes with PostgreSQL
To implement soft deletes straightforwardly, we can use a boolean flag. Let’s implement a straightforward example with the users table.
1 2 3 4 5 |
CREATE TABLE users ( id serial PRIMARY KEY, email text UNIQUE, is_deleted boolean DEFAULT false ) |
For starters, let’s insert a user into our database. Thanks to using the DEFAULT keyword, all our users are not deleted by default.
1 2 3 4 5 |
INSERT into users ( email ) VALUES ( 'marcin@wanago.io' ) |
If we want to perform a soft delete on the above record, we no longer use the DELETE statement. Instead, we perform an UPDATE.
1 2 3 |
UPDATE users SET is_deleted = true WHERE email = 'marcin@wanago.io' |
Now we need to take the is_deleted column into account when performing various other queries. A good example is a query to get the list of all the users.
1 2 |
SELECT * from users WHERE is_deleted = false |
Advantages and disadvantages of soft deletes
The obvious advantage of soft deletes is that we can always easily restore the data we’ve previously deleted. We might think we could also achieve that with frequent backups, but soft deletes could help us achieve a better user experience. For example, we could create an undo button in our application that changes the is_deleted flag to false.
Still, we definitely should create backups and make sure they work.
A significant disadvantage of soft deletes is that we always need to keep them in mind when performing various queries. If we use Object-Relational Mapping such as TypeORM, it makes our work a bit easier. However, we still need to be aware of the additional effort PostgreSQL needs to make to filter out records where is_deleted equals true.
Even if we expect is_deleted to be false in most of our queries, this might not always be the case. A significant advantage of the soft deletes is the fact that we can still fetch the deleted records. This might come in handy when we want to generate some report, for example. Thanks to soft deletes, we can include all of the records.
There is also an important catch in our example with the users table above. When we mark the email column as unique, it works in a predictable way when we decide to use the regular DELETE statements. This is not that straightforward with soft deletes. PostgreSQL takes the records with is_deleted set to true into account when checking if the unique constraint is violated.
Soft deletes can come in handy when dealing with relationships. For example, in this series, we associated posts with users through a many-to-one relationship. Because of that, the posts table has a column called author_id used to save the author of that post. Trying to perform a hard delete on a user that is an author of a post leads to a foreign constraint violation. This does not happen with soft deletes.
We could deal with the above issue using cascade deletes, but this would delete the posts also.
Soft deletes with TypeORM
The TypeORM community anticipated the soft delete feature very much. To start using this feature, we need to use the @DeleteDateColumn decorator.
category.entity.ts
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 |
import { Column, DeleteDateColumn, Entity, ManyToMany, PrimaryGeneratedColumn } from 'typeorm'; import Post from '../posts/post.entity'; @Entity() class Category { @PrimaryGeneratedColumn() public id: number; @Column() public name: string; @ManyToMany(() => Post, (post: Post) => post.categories) public posts: Post[]; @DeleteDateColumn() public deletedAt: Date; } export default Category; |
The above is because TypeORM stores the deletion date instead of a simple boolean flag. Thanks to that, we have an additional piece of information stored that might come in handy.
To perform the soft delete, we need to use the softDelete method instead of delete.
categories.service.ts
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 |
import { Injectable } from '@nestjs/common'; import Category from './category.entity'; import { InjectRepository } from '@nestjs/typeorm'; import { Repository } from 'typeorm'; import CategoryNotFoundException from './exceptions/categoryNotFound.exception'; @Injectable() export default class CategoriesService { constructor( @InjectRepository(Category) private categoriesRepository: Repository<Category> ) {} async deleteCategory(id: number): Promise<void> { const deleteResponse = await this.categoriesRepository.softDelete(id); if (!deleteResponse.affected) { throw new CategoryNotFoundException(id); } } // ... } |
When we do the above, TypeORM sets the deletedAt column to the current date.
If you want to know more about dates in PostgreSQL, check out Managing date and time with PostgreSQL and TypeORM
Performing queries
Thanks to doing the above, we can check if a record has been soft-deleted when performing other queries.
1 2 3 |
getAllCategories(): Promise<Category[]> { return this.categoriesRepository.find({ relations: ['posts'] }); } |
Thankfully, TypeORM does that for us out of the box when we use the find() method. Under the hood, it checks if the deletedAt column has the NULL value.
Still, we might want to allow the users to view a deleted category if they know the id, for example. Our /categories/[id] endpoint by default will respond with 404 Not Found for soft-deleted records, though.
Thankfully, we can explicitly tell TypeORM not to filter out soft-deleted records using the withDeleted argument.
1 2 3 4 5 6 7 8 9 10 11 12 13 |
async getCategoryById(id: number): Promise<Category> { const category = await this.categoriesRepository.findOne( id, { relations: ['posts'], withDeleted: true } ); if (category) { return category; } throw new CategoryNotFoundException(id); } |
When we do the above, we can notice that the deletedAt property is now set.
Restoring deleted records
With soft deletes, we can very easily restore the deleted records. TypeORM allows us to do that with the restore method.
1 2 3 4 5 6 |
async restoreDeletedCategory(id: number) { const restoreResponse = await this.categoriesRepository.restore(id); if (!restoreResponse.affected) { throw new CategoryNotFoundException(id); } } |
When we do the above, TypeORM sets the value of deletedAt back to null.
Summary
In this article, we’ve learned the concept of soft delete. While doing so, we wrote some SQL code. Understanding how the soft delete functions under the hood is essential to grasping its advantages and disadvantages better. Even though TypeORM can help us a lot in implementing soft deletes, it comes with some drawbacks. After going through them in this article, it might appear that we need a pretty good reason to use soft deletes. This is because it is bound to make our database requires more space and take a toll on the performance. Even so, it has its uses and might come in handy in some cases.
Very handy article, thank you!