Data Normalization in Redux — React: Redux Toolkit

Most applications work with data that has a nested structure. For example, blog posts have an author and comments. Comments also have authors and can get likes:

const blogPosts = [
  {
    id: 'post1',
    author: { username: 'user1', name: 'User 1' },
    body: '......',
    comments: [
      {
        id: 'comment1',
        author: { username: 'user2', name: 'User 2' },
        comment: '.....'
      },
      {
        id: 'comment2',
        author: { username: 'user3', name: 'User 3' },
        comment: '.....'
      }
    ]
  },
];

It is hard to work directly with such a structure for several reasons:

• It duplicates data inside it, so it is hard to update
• The logic of reducers gets more complicated the more nesting there is

The correct approach to Redux is to consider it as a relational database. The data inside the repository should be normalized. With this view, we can perceive each slice working with a set of entities as a separate table in the database.

Let's observe the basic principles of organizing data:

• Each entity should be stored in its reducer
• A collection of entities of the same type should be stored as an object, where the keys are object identifiers and the values are the objects themselves
• The order of data in this object should be defined by a separate array consisting only of identifiers
• Data should refer to each other only by identifiers

Now we will discuss that code example:

{
  posts: {
    entities: {
      post1: {
        id: 'post1',
        author: 'user1',
        body: '......',
        comments: ['comment1', 'comment2'],
      },
      post2: {
        id: 'post2',
        author: 'user2',
        body: '......',
        comments: [],
      },
    },
    ids: ['post1', 'post2'],
  },
  comments: {
    entities: {
      comment1: {
        id: 'comment1',
        author: 'user2',
        comment: '.....',
      },
      comment2: {
        id: 'comment2',
        author: 'user3',
        comment: '.....',
      },
    },
    ids: ['comment1', 'comment2'],
  },
  users: {
    entities: {
      user1: {
        id: 'user1',
        username: 'user1',
        name: 'User 1',
      },
      user2: {
        id: 'user2',
        username: 'user2',
        name: 'User 2',
      },
      user3: {
        id: 'user3',
        username: 'user3',
        name: 'User 3',
      },
    },
    ids: ['user1', 'user2', 'user3'],
  }
}

The data are now normalized. We store each entity in its reducer. The entities object stores the entities themselves, and the ids object has the identifiers. What advantages have we gained:

• The data aren't repeated, so if you want to make a change, you only need to do so in one place
• Reducers are not nested
• Data in this form are easy to retrieve and modify

Now let's see what it looks like inside the slices:

const slice = createSlice({
  name: 'users',
  initialState: {
    ids: [],
    entities: {},
  },
  reducers: {
    addUser(state, action) {
      const { user } = action.payload;

      state.entities[user.id] = user;
      state.ids.push(user.id);
    },
    removeUser(state, action) {
      const { userId } = action.payload;

      delete state.entities[userId];
      state.ids = state.ids.filter((id) => id !== userId);
    },
    updateUser(state, action) {
      const { userId, data } = action.payload;

      Object.assign(state.entities[userId], data);
    }
  },
});

dispatch(addUser({ user }));
dispatch(removeUser({ userId }));
dispatch(updateUser({ userId, data }));

normalizr

The data coming from the back end isn't usually normalized, as it is convenient for the front end. Therefore, before adding them to the store, we must normalize them first. We can do it either by manually traversing the collection and converting it to the desired object or by using the normalizr library.

Recommended materials

1. Normalizing State Shape