MongoDB

• No-sql document store
• Stores BSON (Binary json)

Crud

Create

• db.collection.insertOne()
• db.collection.insertMany()

  Update

• db.collection.updateOne(<match on>, <set values>)
• db.collection.updateMany(<match on>, <set values>)
• db.collection.replaceOne(<match on>, <document>)

  Read

• db.collection.find()<.limit(5)>

  Delete

• db.collection.deleteOne()
• db.collection.deleteMany()

There is also a db.collection.bulkWrite() which takes an array of changes i.e. [{insertOne: { ... }}, {updateOne: { ... }}]

Indexing

Indexes are data structures that store a small portion of the collections data in an easy to traverse form.

_id is used as the default index.

• Single field indexes
• Compound indexes
  • Can specify 2 or more indexes
  • Data is grouped by 1st index -> nth index
  • i.e. CustomerId, SystemTypeId, CompanyTypeId
  • Can also define a sort order on these
• Multikey indexes
  • Keys against an array
• Text indexes
  • Good for text search & support text search queries
  • A collection can only have one text index
• Wildcard indexes
  • Indexing fields that we don’t know or are likely to change

Index notes

• You can hide an index instead of deleting it straight away
  • This will stop queries using the index but keep it around so that we can see the impact deleting it will have before actually deleting it
  • db.restaurants.hideIndex({borough: 1, ratings: 1}) or db.restaurants.hideIndex("borough_1_ratings_1")
  • Unhide works in the same way but is unhideIndex rather than hideIndex
• Indexes should be able to fit in memory
  • Measure index use db.orders.aggregate([{$indexStats: { }}])
  • See other indexing stats using db.collection.stats(<option>)
  • Includes index sizes etc
  • Another way to see index sizes are: db.collection.totalIndexSize()

WiredTiger

https://www.mongodb.com/docs/manual/core/wiredtiger/

• MongoDb storage engine since 3.2
• WiredTiger uses document-level concurrency control for write operations. As a result, multiple clients can modify different documents of a collection at the same time.
• MultiVersion Concurrency Control (MVCC). At the start of an operation, WiredTiger provides a point-in-time snapshot of the data to the operation. A snapshot presents a consistent view of the in-memory data.
• Uses a write-ahead log (i.e. journal) in combination with checkpoints to ensure data durability.
  • Persists all data modifications between checkpoints
• Responsible for caching data
  • By default, can use a maximum of 0.5*memory-1GB of memory on the in-memory cache. The rest writes to disk

Read Concern/Write Concern/Read Preference

Replication

• Provide redundancy & high availability
• Should be used in production
• Write to primary database & the primary will replicate this onto the secondary sets
• Provides automatic failover
  • When there is no heartbeat to the primary server there will be a new primary elected
• Can read from a secondary server if specified

Sharding

https://www.mongodb.com/resources/products/capabilities/database-sharding-explained https://www.mongodb.com/docs/manual/sharding/

• Sharding is a method for distributing data across multiple machines.
• Used for horizontal scaling across servers
• Shard keys
  • Single indexed field or compound indexed fields
  • sh.shardCollection()

Aggregation

https://www.mongodb.com/docs/manual/reference/aggregation/

• Aggregation pipeline builds up a set of stages that process documents.
• Each stage preforms an operation on the input documents (filter, group, calculate)
• Once the stage has been processed the results are passed to the next stage
• Can return results for groups of documents (i.e. counts, calculations etc)

example:

db.orders.aggregate( [

   // Stage 1: Filter pizza order documents by pizza size
   {
      $match: { size: "medium" }
   },

   // Stage 2: Group remaining documents by pizza name and calculate total quantity
   {
      $group: { _id: "$name", totalQuantity: { $sum: "$quantity" } }
   }

] )