MongoDB Schema Design | Real World Examples

# CHAPTER 28

Real-World MongoDB Database Design Projects #

1. Introduction #

2. Learning Objectives #

• Architect an E-Commerce system (Products, Carts, Orders).

• Architect a Social Media Platform (Users, Posts, Comments, Follows).

• Architect a Multi-Tenant SaaS application.

• Apply the Extended Reference Pattern (Snapshots).

• Recognize practical scenarios for denormalization.

3. Project 1: The E-Commerce Platform #

// 1. Products Collection (Referenced)
{
  "_id": ObjectId("Prod_Laptop"),
  "name": "Gaming Laptop",
  "live_price": 1200,
  "stock": 45
}

// 2. Shopping Carts Collection (One Cart per User)
// We EMBED the cart items for blazing fast updates ($push / $pull)!
{
  "_id": ObjectId("Cart_User1"),
  "user_id": ObjectId("User_1"),
  "items": [
    { "product_id": ObjectId("Prod_Laptop"), "qty": 1 }
  ]
}

// 3. Orders Collection (The Historical Invoice)
// We use the "Extended Reference Pattern". We Reference the ID, but EMBED the price snapshot!
{
  "_id": ObjectId("Order_99"),
  "user_id": ObjectId("User_1"),
  "status": "Shipped",
  "purchased_items": [
    { 
        "product_id": ObjectId("Prod_Laptop"), 
        "name": "Gaming Laptop", 
        "price_at_checkout": 1200 // CRITICAL: Historical snapshot!
    }
  ]
}

4. Project 2: Social Media Platform (Twitter/X Clone) #

// 1. Users Collection
{
  "_id": ObjectId("User_Alice"),
  "handle": "@alice_codes",
  "follower_count": 1500 // Maintained via $inc operator
}

// 2. Posts Collection
{
  "_id": ObjectId("Post_1"),
  "author_id": ObjectId("User_Alice"), // Referenced
  "content": "MongoDB is awesome!",
  "likes_count": 420,
  "created_at": ISODate("2023-10-01T12:00:00Z")
}

// 3. Comments Collection (Referenced to prevent unbounded array growth!)
{
  "_id": ObjectId("Comment_1"),
  "post_id": ObjectId("Post_1"), // Points back to the Post
  "author_id": ObjectId("User_Bob"),
  "text": "I agree completely."
}

5. Project 3: The Multi-Tenant SaaS App (e.g., Slack or Trello) #

// 1. Organizations Collection (The Tenants)
{
  "_id": ObjectId("Org_TechCorp"),
  "company_name": "TechCorp Inc.",
  "subscription_tier": "Enterprise"
}

// 2. Users Collection
{
  "_id": ObjectId("User_Mike"),
  "org_id": ObjectId("Org_TechCorp"), // Data Isolation Link!
  "email": "mike@techcorp.com",
  "role": "Admin"
}

// 3. Projects Collection
{
  "_id": ObjectId("Proj_Alpha"),
  "org_id": ObjectId("Org_TechCorp"), // Data Isolation Link!
  "name": "Website Redesign",
  "tasks": [ // We can EMBED tasks if the max task count per project is small (< 1000)
    { "title": "Update Logo", "status": "Done" },
    { "title": "Fix CSS", "status": "Pending" }
  ]
}

*(By requiring db.projects.find({ orgid: CurrentUser.orgid }) on every backend query, the Node.js API ensures data never bleeds between companies).*

6. The Danger of Over-Normalization #

7. Mini Project: Auditing the E-Commerce Store #

// The Node.js server listens for price changes
const pipeline = [
  { $match: { "updateDescription.updatedFields.live_price": { $exists: true } } }
];

db.collection("products").watch(pipeline).on("change", async (event) => {
    // Automatically insert an Audit Log document!
    await db.collection("audit_logs").insertOne({
        product_id: event.documentKey._id,
        action: "PRICE_CHANGE",
        timestamp: new Date()
    });
});

8. Common Mistakes #

• Unbounded Arrays in Production: Embedding an array of "Messages" inside a "ChatRoom" document works perfectly in development when there are only 10 messages. Three months into production, the ChatRoom hits 50,000 messages, the document hits 16MB, and the chat room permanently crashes. Always anticipate data growth.

9. Best Practices #

• Draw it First: Never start writing Node.js code or Mongoose schemas until you have physically drawn an Entity Relationship Diagram (ERD) on a whiteboard or piece of paper. Map out the 1:N and N:M links visually. Determine what will be embedded and what will be referenced.

10. Exercises #

1. 1. In the E-Commerce schema, why is there a priceatcheckout field embedded inside the Orders collection?

1. 2. In the Social Media schema, why are Comments placed in their own collection rather than embedded as an array inside the Post document?

11. MongoDB Challenges #

db.projects.find({ org_id: ObjectId("Org_TechCorp") })

12. MCQ Quiz with Answers #

13. Interview Questions #

• Q: Explain the "Extended Reference Pattern" (embedding a snapshot of specific fields while referencing the main document). Give a concrete example of where this is required.

• Q: Walk me through the database schema you would design for a ride-sharing application (like Uber) involving Riders, Drivers, and Trips.

14. FAQs #

15. Summary #

16. Next Chapter Recommendation #