Normalization vs denormalization default in MongoDB - Performance Comparison

Time Complexity: Normalization vs denormalization default

O(n)

Understanding Time Complexity

When working with databases, how data is organized affects how fast queries run.

We want to see how time to get data changes when using normalized or denormalized data in MongoDB.

Scenario Under Consideration

Analyze the time complexity of fetching user orders in two ways: normalized and denormalized.

// Normalized: separate collections
const user = db.users.findOne({ _id: userId });
const orders = db.orders.find({ userId: user._id }).toArray();

// Denormalized: embedded orders
const userWithOrders = db.users.findOne({ _id: userId });
const orders = userWithOrders.orders;
    

This code shows fetching orders separately (normalized) versus embedded inside user (denormalized).

Identify Repeating Operations

• Normalized primary operation: Querying orders collection for matching userId.
• Normalized how many times: Once per user, but scanning orders that belong to user.
• Denormalized primary operation: Single query to users collection, then access embedded orders array.
• Denormalized how many times: One query, no extra scans.

How Execution Grows With Input

As the number of orders grows, how does query time change?

Input Size (orders per user)	Normalized Approx. Operations	Denormalized Approx. Operations
10	Scan 10 orders in orders collection	Access 10 embedded orders
100	Scan 100 orders in orders collection	Access 100 embedded orders
1000	Scan 1000 orders in orders collection	Access 1000 embedded orders

Pattern observation: Both grow roughly linearly with number of orders, but normalized requires separate query scanning orders collection.

Final Time Complexity

Time Complexity: O(n)

This means the time to fetch orders grows linearly with how many orders a user has, whether normalized or denormalized.

Common Mistake

[X] Wrong: "Denormalized data always makes queries faster regardless of data size."

[OK] Correct: Large embedded arrays can slow down queries and updates, so time can still grow with data size.

Interview Connect

Understanding how data layout affects query time helps you design better databases and answer real questions about performance.

Self-Check

"What if we added an index on userId in the orders collection? How would that change the time complexity for normalized queries?"

Practice

1. What is the main advantage of normalization in MongoDB databases?

easy

A. It separates data into collections linked by references for easy updates.

B. It stores all related data together in one document for faster reads.

C. It duplicates data to improve write performance.

D. It automatically creates indexes on all fields.

Solution

1. Step 1: Understand normalization concept

   Normalization means splitting data into separate collections and linking them by references.

2. Step 2: Identify the main benefit

   This separation makes updating data easier because changes happen in one place without duplication.

3. Final Answer:

   It separates data into collections linked by references for easy updates. -> Option A

4. Quick Check:

   Normalization = separate collections + easy updates [OK]

Hint: Normalization means separate collections linked by references [OK]

Common Mistakes:

• Confusing normalization with denormalization
• Thinking normalization duplicates data
• Assuming normalization speeds up reads

2. Which MongoDB document structure shows denormalization?

easy

A. { _id: 1, name: 'Alice' }, { _id: 101, userId: 1, item: 'Book' }

B. { _id: 1, name: 'Alice', orders: [ { orderId: 101, item: 'Book' } ] }

C. { _id: 101, userId: 1, item: 'Book' }

D. { _id: 1, name: 'Alice', orders: null }

Solution

1. Step 1: Identify denormalized structure

   Denormalization stores related data together inside one document, like embedding orders inside user.

2. Step 2: Check options for embedded data

   { _id: 1, name: 'Alice', orders: [ { orderId: 101, item: 'Book' } ] } embeds orders array inside the user document, showing denormalization.

3. Final Answer:

   { _id: 1, name: 'Alice', orders: [ { orderId: 101, item: 'Book' } ] } -> Option B

4. Quick Check:

   Denormalization = embedded related data [OK]

Hint: Denormalization embeds related data inside one document [OK]

Common Mistakes:

• Choosing separate collections as denormalized
• Ignoring embedded arrays as denormalization
• Confusing null fields with embedded data

3. Given these two collections:
users: { _id: 1, name: 'Bob' }
orders: { _id: 101, userId: 1, item: 'Pen' }
What is the main drawback of this normalized design when reading user orders?

medium

A. It requires multiple queries or a join-like operation to get all orders for a user.

B. It duplicates order data inside each user document.

C. It stores all orders inside the user document causing large documents.

D. It prevents updating user names easily.

Solution

1. Step 1: Understand normalized design

   Users and orders are in separate collections linked by userId reference.

2. Step 2: Identify drawback when reading

   To get all orders for a user, you must query orders collection filtering by userId, requiring multiple queries or aggregation.

3. Final Answer:

   It requires multiple queries or a join-like operation to get all orders for a user. -> Option A

4. Quick Check:

   Normalized read = multiple queries [OK]

Hint: Normalized data needs multiple queries to combine related info [OK]

Common Mistakes:

• Thinking normalized data duplicates info
• Assuming all data is embedded in one document
• Believing updates are harder in normalized data

4. You have a denormalized MongoDB document:
{ _id: 1, name: 'Carol', orders: [ { orderId: 201, item: 'Notebook' } ] }
Which problem can occur if you update the item name in one order but forget to update it elsewhere?

medium

A. Query performance slows down because of references.

B. Indexes on orders array are lost.

C. The database schema becomes normalized automatically.

D. Data inconsistency due to duplicated order info in multiple documents.

Solution

1. Step 1: Recognize denormalization risk

   Denormalization duplicates related data inside documents, so the same order info may appear in many places.

2. Step 2: Understand update problem

   If you update one copy but not others, data becomes inconsistent and unreliable.

3. Final Answer:

   Data inconsistency due to duplicated order info in multiple documents. -> Option D

4. Quick Check:

   Denormalization risk = data inconsistency [OK]

Hint: Denormalization can cause inconsistent duplicated data if not updated everywhere [OK]

Common Mistakes:

• Thinking denormalization slows queries
• Believing schema changes automatically
• Confusing index loss with denormalization

5. You want to design a MongoDB schema for a blog with users and posts.
Users have many posts, and posts rarely change after creation.
Which design is best for fast reading and why?

Options:
A: Store users and posts in separate collections (normalized).
B: Embed all posts inside each user document (denormalized).
C: Duplicate posts in both users and posts collections.
D: Store posts only, with user info duplicated in each post.

hard

A. Separate collections for users and posts for easy updates.

B. Store posts only with duplicated user info for simpler queries.

C. Embed posts inside user documents for fast reads since posts rarely change.

D. Duplicate posts in both collections to optimize writes.

Solution

1. Step 1: Analyze data change frequency

   Posts rarely change, so embedding them inside users won't cause frequent update problems.

2. Step 2: Choose design for fast reads

   Embedding posts inside user documents allows fetching user and posts in one read, improving read speed.

3. Step 3: Compare options

   Embedding posts inside user documents for fast reads since posts rarely change fits best for fast reads with rare updates; separate collections require joins; duplicating posts in both risks inconsistency; storing posts only duplicates user info unnecessarily.

4. Final Answer:

   Embed posts inside user documents for fast reads since posts rarely change. -> Option C

5. Quick Check:

   Denormalization + rare updates = embed for fast reads [OK]

Hint: Embed rarely changing related data for faster reads [OK]

Common Mistakes:

• Choosing normalization for fast reads
• Duplicating data causing inconsistency
• Ignoring update frequency in design