Collection Organization

📦 NoSQL Collection Organization

Your Data’s Home: Collection Design, Capped Collections & Views

🏠 The Big Picture: What Are Collections?

Imagine your bedroom. You have different containers to organize your stuff:

• A toy box for toys
• A bookshelf for books
• A drawer for clothes

In NoSQL databases (like MongoDB), collections are like these containers. Each collection holds documents (your data items) that belong together.

But here’s the magical part: Unlike a toy box with a fixed shape, collections in NoSQL are flexible—you can put different shaped toys in the same box!

📐 Collection Design

What Is Collection Design?

Collection design is deciding how to organize your data containers. It’s like deciding:

• Should toys and books go in the same box?
• Or should they have separate homes?

The Golden Rule: Group Related Things Together

graph TD
    A["Your Data"] --> B["Users Collection"]
    A --> C["Products Collection"]
    A --> D["Orders Collection"]
    B --> B1["User 1"]
    B --> B2["User 2"]
    C --> C1["Product A"]
    C --> C2["Product B"]

Two Main Approaches

1. Embedding (Putting Things Inside)

Like putting a letter INSIDE an envelope:

{
  "name": "Sarah",
  "address": {
    "street": "123 Main St",
    "city": "Boston"
  }
}

When to embed:

• Data is always accessed together
• Child data is unique to parent
• Small amounts of related data

2. Referencing (Pointing to Another Place)

Like writing “See toy box #5” instead of putting the toy here:

{
  "name": "Sarah",
  "address_id": "addr_12345"
}

When to reference:

• Data is shared across documents
• Data changes frequently
• Large amounts of related data

Design Questions to Ask

Real Example: Blog Posts

Option A - Embedded Comments:

{
  "title": "My First Post",
  "comments": [
    {"user": "Tom", "text": "Great!"},
    {"user": "Amy", "text": "Love it!"}
  ]
}

Option B - Referenced Comments:

{
  "title": "My First Post",
  "comment_ids": ["c1", "c2", "c3"]
}

Which is better? It depends!

• Few comments that you always show? → Embed
• Thousands of comments with pagination? → Reference

🎪 Capped Collections

What Are Capped Collections?

Imagine a circular toy train track. The train goes round and round. When you add a new train car at the front, the oldest one at the back gets pushed off!

Capped collections work exactly like this:

• Fixed size (you decide: “only 1000 documents” or “only 5MB”)
• Oldest documents automatically deleted when full
• New documents always go at the end
• SUPER fast for inserting and reading in order!

graph LR
    subgraph Capped Collection
    A["Newest"] --> B["New"] --> C["Old"] --> D["Oldest"]
    end
    E["New Doc"] -.-> A
    D -.-> F["Auto Deleted"]

Creating a Capped Collection

db.createCollection("logs", {
  capped: true,
  size: 5242880,    // 5 MB max
  max: 5000         // 5000 documents max
})

Two limits work together:

• size: Maximum bytes (required)
• max: Maximum document count (optional)

Whichever limit is hit first triggers deletion!

Perfect Use Cases

Capped Collection Rules

Can DO:

• Insert new documents ✅
• Read documents ✅
• Update documents (if size stays same) ✅

Cannot DO:

• Delete individual documents ❌
• Update to make document bigger ❌
• Shard the collection ❌

Example: Chat Room Messages

// Create capped collection for chat
db.createCollection("chat_room_1", {
  capped: true,
  size: 1048576,  // 1 MB
  max: 500        // Last 500 messages
})

// Insert a message
db.chat_room_1.insertOne({
  user: "Alex",
  message: "Hello everyone!",
  timestamp: new Date()
})

// Get messages in order (super fast!)
db.chat_room_1.find().sort({$natural: 1})

👁️ Views

What Are Views?

Think of views as magic windows that show you a specific part of your data—filtered, transformed, or combined—without copying anything!

It’s like having a window in your room that ONLY shows red toys, even though your toy box has all colors.

graph TD
    A["Original Collection"] --> B["View: Active Users"]
    A --> C["View: Recent Orders"]
    A --> D["View: VIP Customers"]
    B --> B1["Shows only<br>active users"]
    C --> C1["Shows orders<br>from last 7 days"]
    D --> D1["Shows users<br>spending > $1000"]

Why Use Views?

Creating a View

Scenario: You have a users collection but want a view showing only active users.

db.createView(
  "active_users",     // View name
  "users",            // Source collection
  [                   // Pipeline (filters)
    { $match: { status: "active" } },
    { $project: {
        name: 1,
        email: 1,
        _id: 0
      }
    }
  ]
)

Now active_users acts like a collection:

db.active_users.find()
// Returns only active users!

View vs Collection

Real Examples

Example 1: Orders This Week

db.createView(
  "recent_orders",
  "orders",
  [
    { $match: {
        orderDate: {
          $gte: new Date(Date.now() - 7*24*60*60*1000)
        }
      }
    }
  ]
)

Example 2: Product Summary (Join)

db.createView(
  "product_details",
  "products",
  [
    { $lookup: {
        from: "categories",
        localField: "category_id",
        foreignField: "_id",
        as: "category"
      }
    },
    { $unwind: "$category" },
    { $project: {
        name: 1,
        price: 1,
        categoryName: "$category.name"
      }
    }
  ]
)

Modifying Views

Views can be modified but not directly edited:

// Drop and recreate
db.active_users.drop()
db.createView("active_users", "users", [...])

// Or use collMod
db.runCommand({
  collMod: "active_users",
  viewOn: "users",
  pipeline: [...]
})

🎯 Quick Decision Guide

graph TD
    A["What do you need?"] --> B{Store actual data?}
    B -->|Yes| C{Fixed size buffer?}
    B -->|No| D["Use a VIEW"]
    C -->|Yes| E["Use CAPPED COLLECTION"]
    C -->|No| F["Use REGULAR COLLECTION"]
    F --> G{Related data?}
    G -->|Always together| H["EMBED"]
    G -->|Shared/Large| I["REFERENCE"]

🌟 Summary: Your Data Organization Toolbox

You did it! Now you know how to:

• Design collections that make sense
• Use capped collections for logs and recent data
• Create views for filtered, secure access

Your data has a happy, organized home! 🏠✨