Understanding Collections

Introduction

A collection is a group of documents — the MongoDB analogue of a SQL table, but without an enforced schema. Collections live inside databases, and databases live inside the MongoDB cluster.

Understanding the topic

Collection essentials:

• Collections are created on first insert — no CREATE TABLE required.
• Documents in a collection don't need identical schemas, but in practice you keep them similar.
• Use one collection per logical entity: users, orders, products.
• Each collection has its own indexes, validation rules, and storage stats.
• Capped collections, time-series collections and clustered collections are specialized variants.

Syntax reference

// Create implicitly
db.products.insertOne({ name: "Mug" });

// Or explicitly with validation
db.createCollection("orders", {
  validator: {
    $jsonSchema: {
      required: ["userId", "total"],
      properties: {
        userId: { bsonType: "objectId" },
        total:  { bsonType: "decimal", minimum: 0 }
      }
    }
  }
});

Real-world use

A typical SaaS app might have collections for users, workspaces, projects, events, auditLogs, billing. Each is independently indexed and queried.

Best practices

• Name collections in lowercase, plural: users not User.
• Add $jsonSchema validators when the shape stabilizes.
• Use time-series collections for IoT and metrics — they're 10× more space-efficient.