MongoDB — Basics

MongoDB — Basics

What it is

• Document database. Stores BSON (binary JSON).
• Flexible schema — no enforced structure unless you add validators.
• Horizontally scalable via sharding.

Document model

• Database → Collections → Documents (BSON).
• Documents up to 16MB.
• Nested documents, arrays, ObjectId, Decimal128, ISODate, UUID.

CRUD

db.users.insertOne({ name: 'A', email: 'a@b.com' })
db.users.insertMany([{...},{...}])

db.users.find({ age: { $gte: 18 } })
db.users.findOne({ _id: ObjectId('...') })

db.users.updateOne({ _id }, { $set: { name: 'B' }, $inc: { logins: 1 } })
db.users.updateMany({ active: false }, { $set: { archived: true } })

db.users.deleteOne({ _id })
db.users.deleteMany({ status: 'dead' })

db.users.replaceOne({ _id }, { ...full doc... })

Operators

• Comparison: $eq, $ne, $gt, $gte, $lt, $lte, $in, $nin.
• Logical: $and, $or, $not, $nor.
• Element: $exists, $type.
• Array: $all, $elemMatch, $size.
• Update: $set, $unset, $inc, $push, $pull, $addToSet, $pop, $rename.

Indexes

• Default _id index. Add others.
• Types:
  • Single field: db.users.createIndex({ email: 1 }) (1 asc, -1 desc).
  • Compound: {a: 1, b: -1}. Leftmost prefix matters.
  • Multikey — auto when indexed field is array.
  • Text: db.posts.createIndex({ body: 'text' }) for full-text search (basic).
  • 2dsphere — geospatial.
  • Hashed — for sharding.
  • TTL — auto-expire docs: createIndex({ createdAt: 1 }, { expireAfterSeconds: 3600 }).
  • Unique, partial, sparse, wildcard.
• Use .explain('executionStats') to verify index use.

Aggregation pipeline

db.orders.aggregate([
  { $match: { status: 'paid' } },
  { $group: { _id: '$user', total: { $sum: '$amount' } } },
  { $sort: { total: -1 } },
  { $limit: 10 }
])

Stages: $match, $project, $group, $sort, $limit, $skip, $unwind, $lookup (join), $facet, $bucket, $addFields, $out, $merge.

Transactions

• Multi-document transactions on replica sets (4.0+) and sharded clusters (4.2+).
• Performance overhead — use sparingly. Prefer single-document atomic ops where possible.

Replica sets

• Primary + secondaries + optional arbiter.
• Async replication via oplog.
• Automatic failover (election) on primary loss.
• Read preference: primary, primaryPreferred, secondary, secondaryPreferred, nearest.
• Write concern: w: 1 (primary), w: 'majority' (durable).
• Read concern: local, available, majority, snapshot.

Sharding

• Horizontal partitioning across shards (each is a replica set).
• Shard key determines distribution. Choose carefully: high cardinality, low frequency, monotonic NOT good (hotspot last shard).
• Hashed sharding for even distribution; ranged for query locality.
• Config servers store metadata. mongos routes queries.

Tools

• Driver: official MongoDB driver (Node, Python pymongo/motor, Go, Java).
• ODM: Mongoose (Node), MongoEngine (Python).
• CLI: mongosh. GUI: Compass.