nosql-7.md

07 — Quick Rules of Thumb

The golden rules for deciding how to structure your MongoDB data.

---

The Big Decision: Embed vs. Reference

Every time you have related data, ask yourself one question:

Does this data belong to ONE parent, or is it SHARED across many?

Situation	Strategy	How
Data belongs to one parent	Embed it	Put it inside the document
Data is shared across many	Reference it	Use $lookup

---

✅ Embed When...

Rule: The data has one owner and is always read together with it.

Example: Comments on a blog post.

{
  "_id": "post_1",
  "title": "My First Post",
  "comments": [
    { "author": "Alice", "text": "Great post!" },
    { "author": "Bob",   "text": "Thanks for sharing." }
  ]
}

Why embed?

• Comments are meaningless without the post.
• You always read comments at the same time as the post.
• No other document needs to reference the same comment.

---

🔗 Reference When...

Rule: The data is shared or reused across multiple documents.

Example: A product used in many orders.

// Orders collection
{ "_id": "order_1", "product_id": "prod_42", "qty": 2 }
{ "_id": "order_2", "product_id": "prod_42", "qty": 1 }

// Products collection (separate)
{ "_id": "prod_42", "name": "Keyboard", "price": 79.99 }

Why reference?

• The product is shared across many orders.
• If the price changes, you update it in one place.
• Embedding would duplicate the product data everywhere.

---

🚩 The Red Flag

If you are using $lookup for every single query, you are likely trying to build a SQL database in NoSQL.

This is the most common MongoDB mistake. If your queries always look like this:

// Bad sign — doing this for EVERY query
db.orders.aggregate([
  { $lookup: { from: "customers", ... } },
  { $lookup: { from: "products", ... } },
  { $lookup: { from: "discounts", ... } }
])

Consider:

1. Embedding more data directly into documents.
2. Or switching to a relational database (SQL) if your data is truly relational.

---

Decision Flowchart

Is the related data accessed together with the parent document?
│
├── YES → Is it owned by only ONE parent?
│           ├── YES → EMBED IT ✅
│           └── NO  → REFERENCE IT 🔗
│
└── NO  → Is it a small, rarely-changing lookup?
            ├── YES → EMBED A SNAPSHOT (e.g., store product name at time of order)
            └── NO  → REFERENCE IT 🔗

---

Quick Reference Summary

Concept	SQL	MongoDB
Store related data	Foreign key + JOIN table	Embed array in document
Shared/reusable data	Foreign key	$lookup reference
Query an array	Complex JOIN	{ field: "value" }
Filter	WHERE	$match / .find({ })
Select columns	SELECT col1, col2	{ col1: 1, col2: 1 }
Aggregate	GROUP BY + SUM()	$group + $sum
Sort	ORDER BY	$sort
Limit rows	LIMIT n	$limit: n