5

API with NestJS #75. Many-to-many relationships using raw SQL queries

 2 years ago
source link: https://wanago.io/2022/09/19/api-nestjs-many-to-many-relationships-raw-sql/
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.
neoserver,ios ssh client

API with NestJS #75. Many-to-many relationships using raw SQL queries

JavaScript NestJS SQL

September 19, 2022

This entry is part 75 of 75 in the API with NestJS

Designing relationships between tables is one of the crucial parts of working with databases. In this article, we look into a more complex relationship called many-to-many.

You can find the code from this article in this repository.

The many-to-many relationship

A many-to-many relationship happens when many records in one table relate to many records in another table. A good example is a connection between posts and categories. A particular post can be published under multiple categories. For example, this article falls under both the SQL and JavaScript categories. On the other hand, a single category can be related to numerous different posts.

So far, we’ve worked with one-to-one or many-to-one relationships using raw SQL queries. In the above approaches, we use a simple column containing a foreign key that matches a row from a different table.

one-to-many-categories.png

The case gets complicated when we want to create a connection between one post and multiple categories. We shouldn’t put multiple values in the category_id column. To implement a many-to-many relationship, we create a joining table.

categories_posts.png

Creating the categories_posts table allows us to store the relationships between particular categories and posts.

Creating the many-to-many relationship

Let’s define a migration that creates the categories and categories_posts tables.

npx knex migrate:make add_categories_table
20220914233800_add_categories_table.ts
import { Knex } from 'knex';
export async function up(knex: Knex): Promise<void> {
  return knex.raw(`
    CREATE TABLE categories (
      id int GENERATED BY DEFAULT AS IDENTITY PRIMARY KEY,
      name text NOT NULL
    CREATE TABLE categories_posts (
      category_id int REFERENCES categories(id),
      post_id int REFERENCES posts(id),
      PRIMARY KEY (category_id, post_id)
export async function down(knex: Knex): Promise<void> {
  return knex.raw(`
    DROP TABLE categories, categories_posts;

An important thing to notice in how we created the categories_posts table is that it doesn’t have a separate id column. Instead, we specify a composite primary key. This approach has some advantages. First, we save a bit of disk space thanks to not creating the id column. But more importantly, we make sure it is unique thanks to marking a combination of the category_id and post_id as the primary key. All rows in a table should have a different primary key. Thanks to that, the following data would never appear in our table:

category-data.png

By the above, we ensure that a particular post might relate to a particular category only once.

Connecting posts to categories

When a user publishes a post, it can be related to multiple categories. For example, we might accept the following data through our API:

  "title": "My first post",
  "content": "Hello world!",
  "categoryIds": [1, 2]

The above means that we want to add two rows to the categories_posts table:

category-data-2.png

Fortunately, we can insert multiple rows into a table simultaneously. One way of doing that is inserting a result of a SELECT  query:

INSERT INTO categories_posts (
  post_id, category_id
  SELECT 1 as post_id, unnest(ARRAY[1,2]) AS category_id
  FROM created_post

To understand the above code, we need to take a closer look at this SELECT query:

SELECT 1 as post_id, unnest(ARRAY[1,2]) AS category_id

category-data-4.png

Above, we use the unnest function to expand an array to a set of rows. Thanks to that, our SELECT query returns multiple rows that the INSERT statement saves into the database.

We can now use all of the above knowledge to create a post and connect it to categories in the same query.

posts.repository.ts
import { Injectable } from '@nestjs/common';
import DatabaseService from '../database/database.service';
import PostDto from './post.dto';
import PostWithCategoryIdsModel from './postWithCategoryIds.model';
@Injectable()
class PostsRepository {
  constructor(private readonly databaseService: DatabaseService) {}
  async createWithCategories(postData: PostDto, authorId: number) {
    const databaseResponse = await this.databaseService.runQuery(
      WITH created_post AS (
        INSERT INTO posts (
          title,
          post_content,
          author_id
        ) VALUES (
        ) RETURNING *
      created_relationships AS (
        INSERT INTO categories_posts (
          post_id, category_id
          SELECT created_post.id AS post_id, unnest($4::int[]) AS category_id
          FROM created_post
      SELECT *, $4 as category_ids FROM created_post
      [postData.title, postData.content, authorId, postData.categoryIds],
    return new PostWithCategoryIdsModel(databaseResponse.rows[0]);
  // ...
export default PostsRepository;

We also need to create a model that includes the categoryIds property.

postWithCategoryIds.model.ts
import PostModel, { PostModelData } from './post.model';
interface PostWithCategoryIdsModelData extends PostModelData {
  category_ids: number[] | null;
class PostWithCategoryIdsModel extends PostModel {
  categoryIds: number[];
  constructor(postData: PostWithCategoryIdsModelData) {
    super(postData);
    this.categoryIds = postData.category_ids || [];
export default PostWithCategoryIdsModel;

Thanks to the above, we can now create posts and connect them to categories in a single query.

Screenshot-from-2022-09-15-23-40-35.png

Fetching the ids of categories of a certain post

So far, when fetching the details of a certain post, we’ve attached the details of an author. Let’s take it a step further, and attach the ids of the categories related to the post. Let’s break down this problem into a simple set of steps to perform.

First, we need to get all the rows from the categories_posts table related to a particular post.

SELECT category_id FROM categories_posts
WHERE post_id = 3

category-data-7.png

We can parse it into a single array to make it easier to work with.

SELECT ARRAY(
  SELECT category_id FROM categories_posts
  WHERE post_id = 3
) AS category_ids

category-data-8.png

Let’s prepare a new model to handle the above data.

postWithDetails.model.ts
import PostModel, { PostModelData } from './post.model';
import UserModel from '../users/user.model';
interface PostWithDetailsModelData extends PostModelData {
  user_id: number;
  user_email: string;
  user_name: string;
  user_password: string;
  address_id: number | null;
  address_street: string | null;
  address_city: string | null;
  address_country: string | null;
  category_ids: number[] | null;
class PostWithDetails extends PostModel {
  author: UserModel;
  categoryIds: number[];
  constructor(postData: PostWithDetailsModelData) {
    super(postData);
    this.author = new UserModel({
      ...postData,
      id: postData.user_id,
      email: postData.user_email,
      name: postData.user_name,
      password: postData.user_password,
    this.categoryIds = postData.category_ids || [];
export default PostWithDetails;

We now have everything we need to fetch a post with its author and category ids.

posts.repository.ts
import { Injectable, NotFoundException } from '@nestjs/common';
import DatabaseService from '../database/database.service';
import PostWithDetails from './postWithDetails.model';
@Injectable()
class PostsRepository {
  constructor(private readonly databaseService: DatabaseService) {}
  async getWithDetails(postId: number) {
    const postResponse = await this.databaseService.runQuery(
      SELECT
        posts.id AS id, posts.title AS title, posts.post_content AS post_content, posts.author_id as author_id,
        users.id AS user_id, users.email AS user_email, users.name AS user_name, users.password AS user_password,
        addresses.id AS address_id, addresses.street AS address_street, addresses.city AS address_city, addresses.country AS address_country
      FROM posts
      JOIN users ON posts.author_id = users.id
      LEFT JOIN addresses ON users.address_id = addresses.id
      WHERE posts.id=$1
      [postId],
    const postEntity = postResponse.rows[0];
    if (!postEntity) {
      throw new NotFoundException();
    const categoryIdsResponse = await this.databaseService.runQuery(
      SELECT ARRAY(
        SELECT category_id FROM categories_posts
        WHERE post_id = $1
      ) AS category_ids
      [postId],
    return new PostWithDetails({
      ...postEntity,
      category_ids: categoryIdsResponse.rows[0].category_ids,
  // ...
export default PostsRepository;
Screenshot-from-2022-09-16-02-15-52.png

Fetching all posts from a certain category

There is a big chance that we will want to get a list of all the posts from a certain category. To achieve this, we need to join the data from the posts table with categories_posts.

Let’s break down this problem into smaller chunks. First, we must fetch all post ids from a certain category.

SELECT post_id
FROM categories_posts
WHERE category_id = 1

category-data-5.png

Since we know the ids of all the posts, we can use the JOIN  statement to match them with the rows from the posts table.

SELECT posts.id AS post_id, posts.title AS post_title, posts.post_content AS post_content, posts.author_id AS author_id
FROM categories_posts
JOIN posts ON posts.id=categories_posts.post_id
WHERE category_id = 1
category-data-6.png

Let’s create a new model to prepare for the above data.

categoryWithPosts.model.ts
import CategoryModel, { CategoryModelData } from './category.model';
import PostModel, { PostModelData } from '../posts/post.model';
export interface CategoryWithPostsModelData extends CategoryModelData {
  posts: PostModelData[];
class CategoryWithPostsModel extends CategoryModel {
  posts: PostModel[];
  constructor(categoryData: CategoryWithPostsModelData) {
    super(categoryData);
    this.posts = categoryData.posts.map((postData) => {
      return new PostModel(postData);
export default CategoryWithPostsModel;

We now can use all of the above to:

  • fetch the data of a particular category,
  • match it with related posts,
  • fit the data into the model.
import { Injectable, NotFoundException } from '@nestjs/common';
import DatabaseService from '../database/database.service';
import CategoryWithPostsModel from './categoryWithPosts.model';
@Injectable()
class CategoriesRepository {
  constructor(private readonly databaseService: DatabaseService) {}
  async getCategoryWithPosts(categoryId: number) {
    const categoriesDatabaseResponse = await this.databaseService.runQuery(
      SELECT * FROM categories WHERE id=$1
      [categoryId],
    if (!categoriesDatabaseResponse.rows[0]) {
      throw new NotFoundException();
    const postsDatabaseResponse = await this.databaseService.runQuery(
      SELECT posts.id AS id, posts.title AS title, posts.post_content AS post_content, posts.author_id AS author_id
      FROM categories_posts
      JOIN posts ON posts.id=categories_posts.post_id
      WHERE category_id = $1
      [categoryId],
    return new CategoryWithPostsModel({
      ...categoriesDatabaseResponse.rows[0],
      posts: postsDatabaseResponse.rows,
  // ...
export default CategoriesRepository;

We can use the above logic to fetch the details of the category when it is requested.

Screenshot-from-2022-09-16-01-02-04.png

Summary

In this article, we’ve gone through the many-to-many relationship. When doing that, we implemented an example with posts and categories. To do that, we learned how to manage a joining table and insert multiple entities into the database with one query. There is still more to learn about using NestJS with raw SQL queries, so stay tuned!

Series Navigation<< API with NestJS #74. Designing many-to-one relationships using raw SQL queries


About Joyk


Aggregate valuable and interesting links.
Joyk means Joy of geeK