What is state normalization?

Question

Accepted Answer

**Normalization** means storing data in a flat, relational structure — keyed by **id** in lookup tables — instead of deeply nested objects/arrays. It's the same idea as database normalization, applied to client state, and it makes updates and lookups efficient and consistent.

## The problem with nested/denormalized state

```js
// ❌ nested & duplicated — the same user appears in multiple posts
posts: [
  { id: 1, title: "A", author: { id: 10, name: "Ann" } },
  { id: 2, title: "B", author: { id: 10, name: "Ann" } }, // Ann duplicated
]
// To rename Ann, you must find and update EVERY copy → easy to miss one (drift)
// To find post #2, you must scan the array
```

Duplicated nested data must be updated in many places (inconsistency risk), and finding an item means scanning arrays (O(n)).

## The normalized shape

```js
// ✅ flat lookup tables keyed by id, with references by id
{
  posts: {
    byId: { 1: { id: 1, title: "A", authorId: 10 },   // reference the author by id
            2: { id: 2, title: "B", authorId: 10 } },
    allIds: [1, 2],                                     // order/iteration
  },
  users: {
    byId: { 10: { id: 10, name: "Ann" } },             // each user stored ONCE
  },
}
```

Each entity is stored once, keyed by id; relationships are just id references.

## The benefits

```js
state.users.byId[10];                        // ✅ O(1) lookup by id
state.users.byId[10] = { ...u, name: "Bob" }; // ✅ update Ann ONCE → reflected everywhere
state.posts.allIds.map(id => state.posts.byId[id]); // iterate via the id list
```

```text
✓ Single source of truth per entity → no duplication, no drift
✓ O(1) lookups/updates by id (vs scanning arrays)
✓ Easy to update one entity and have all references reflect it
✓ Simpler to manage relationships and merge new data
```

## Tools

```text
Redux Toolkit's createEntityAdapter generates normalized { ids, entities }
structures + CRUD reducers automatically. normalizr can normalize API responses.
```

## Why it matters

Normalization is the standard pattern for managing **relational/collection data** in client state (especially in Redux).

It enforces single-source-of-truth per entity (update once, reflected everywhere), gives fast id-based lookups, and avoids the bugs and performance issues of deeply nested, duplicated data.

For apps with lists of related entities (users, posts, comments) that get updated and cross-referenced, normalizing the store — often via `createEntityAdapter` — is a key technique for keeping state consistent and efficient.