MongoDB Fundamentals

After this lesson, you will be able to: Use the MongoDB shell: databases, collections, documents, BSON types.

Mongo's data model is simple once you see it. Databases hold collections; collections hold documents; documents are BSON (binary JSON).

Prerequisites:Document Databases vs Relational

The hierarchy

Cluster → Databases → Collections → Documents. Database ≈ database in SQL. Collection ≈ table in SQL (but schemaless by default). Document ≈ row in SQL (but with nested structure).

Shell basics

mongosh commands you'll use daily.

tsx

// Connect, Atlas connection string or local
mongosh "mongodb+srv://user:[email protected]"

// Server info
test> show dbs
test> use myapp                     // switch to (or create) database 'myapp'
myapp> show collections

// Create a doc
myapp> db.users.insertOne({
  name: "Alex",
  email: "[email protected]",
  age: 30,
  tags: ["admin", "beta"],
  address: { city: "NYC", zip: "10001" },
  createdAt: new Date()
})

// Read
myapp> db.users.find()
myapp> db.users.find().pretty()
myapp> db.users.findOne({ name: "Alex" })

// Delete the test db
myapp> db.dropDatabase()

BSON types — beyond JSON

BSON adds types JSON doesn't have.

tsx

{
  _id: ObjectId("507f1f77bcf86cd799439011"),   // Mongo's default ID type (12 bytes)
  email: "[email protected]",                              // String
  age: 30,                                       // Int32
  views: NumberLong("9999999999"),               // Int64 (Long)
  balance: NumberDecimal("49.99"),               // Decimal128, use for money
  active: true,                                  // Boolean
  tags: ["a", "b"],                              // Array
  nested: { foo: 1 },                            // Embedded document
  createdAt: ISODate("2026-01-01T00:00:00Z"),    // Date
  pic: BinData(0, "base64payload"),              // Binary
  ref: DBRef("users", ObjectId("...")),          // Reference (rare)
  pos: { type: "Point", coordinates: [-73, 40] } // GeoJSON
}

Indexes (preview — full lesson later)

Mongo indexes look familiar.

tsx

db.users.createIndex({ email: 1 }, { unique: true });
db.users.createIndex({ "address.city": 1, age: -1 });   // compound
db.users.getIndexes();

Common mistakes

Forgetting BSON has types JSON doesn't, storing money as plain number (Double) → rounding bugs. Treating _id as random ID when it encodes a timestamp, useful for sort-by-creation. Forgetting unique index on email → duplicates pile up. Inserting unbounded nested arrays (e.g., 'all messages in conversation'), doc size limit is 16 MB.

←Document vs Relational Databases

Back to MongoDB

CRUD Operations→