CRUD & Collections

Concept#

Document-Oriented NoSQL#

In relational databases, data is split into structured rows and tables, requiring foreign keys to join relationships.

MongoDB stores data as documents (BSON format, a binary representation of JSON) grouped inside collections. Documents can have varying structures, nested objects, and arrays:

{
  "_id": "60c72b2f9b1d8b2bad000001",
  "name": "Ada Lovelace",
  "email": "ada@example.com",
  "skills": ["math", "programming"],
  "profile": {
    "title": "First Programmer",
    "verified": true
  }
}

Basic CRUD Operations#

• Create: Insert documents into a collection.

  db.users.insertOne({ name: "Grace Hopper", email: "grace@example.com" });

• Read: Query documents using query filters.

  db.users.find({ "profile.verified": true });

• Update: Modify existing documents using update operators like $set, $push, or $inc.

Common Mistakes#

1. Forgetting to use update operators like $set#

If you run db.users.updateOne({ id: 1 }, { status: "active" }), MongoDB replaces the entire document with { status: "active" } instead of only updating the status field. Always wrap modifications in $set or other operators:

db.users.updateOne({ id: 1 }, { $set: { status: "active" } });

2. Treating MongoDB like a relational database#

Creating excessive joins (via $lookup) in MongoDB is slow because NoSQL databases are optimized for denormalized, nested data structures. Embed documents unless data relations change frequently.

Best Practices#

• De-normalize by default: Nest data (like user addresses) inside the parent document if the data is read together and does not grow infinitely.
• Always limit queries: Never run db.collection.find() on large production tables without .limit(n) offsets.
• Use index projection: Select only the fields you need using .find({}, { name: 1, email: 1 }) to reduce memory and transfer bandwidth.