MongoDB Intro

// ===== The model =====
// - Database -> Collection -> Document (BSON)
// - Documents are JSON-like; can nest objects + arrays
// - Schema is enforced by the app, not the database (with optional schema validation)
// - Queries are JSON expressions evaluated server-side

// ===== Hello, collection =====
use shop;
db.orders.insertOne({
  customer: { id: 'c-1', name: 'Alex Chen' },
  total_cents: 4995,
  lines: [{ sku: 'A-1', qty: 2 }, { sku: 'B-2', qty: 1 }],
  tags: ['retail', 'shipped'],
  created_at: new Date(),
});

// ===== Read =====
db.orders.findOne({ 'customer.id': 'c-1' });
db.orders.find({ total_cents: { $gte: 1000 } }).sort({ created_at: -1 }).limit(20);

// ===== Update =====
db.orders.updateOne(
  { _id: id },
  { $set: { 'customer.email': 'a@x.io' }, $push: { tags: 'vip' } }
);

// ===== Aggregate =====
db.orders.aggregate([
  { $match: { 'customer.id': 'c-1' } },
  { $group: { _id: null, total: { $sum: '$total_cents' }, n: { $sum: 1 } } },
]);

// ===== Indexes =====
db.orders.createIndex({ 'customer.id': 1, created_at: -1 });
db.orders.createIndex({ tags: 1 });          // multikey
db.orders.createIndex({ '$**': 'text' });    // full-text

// ===== Replica sets + sharding =====
// Replica set: primary + secondaries; secondaries serve reads with readPreference
// Sharded cluster: data partitioned by shard key across shards; mongos routes queries

// ===== When MongoDB wins =====
// - Variable / sparse document shapes
// - Hierarchical data that maps cleanly to JSON
// - Fast development iteration (no migrations for shape changes)
// - Geospatial / time series with built-in indexes

// ===== When MongoDB hurts =====
// - Strict relational integrity across many tables
// - Heavy multi-document transactions (supported but costlier than Postgres)
// - Reporting / BI workloads (use a warehouse or analytic store)

// ===== Patterns to internalise =====
// - Pre-aggregate documents to match read patterns (denormalise on purpose)
// - Index BEFORE you scale; explain() any new query
// - Schema validation when teams grow; collMod $jsonSchema
// - Use replica set readPreference for read scale; not as a sharding substitute

// ===== Pitfalls =====
// - Treating Mongo like Postgres (joins everywhere via $lookup)
// - Unbounded arrays inside documents -> 16MB cap, slow updates
// - No index on the query you run -> COLLSCAN
// - Reading from secondaries without understanding staleness

// MongoDB stores JSON-shaped documents in collections.
db.customers.insertOne({ name: 'Ada', age: 36 });