UoL/CM3010 Databases and Advanced Data Techniques/Week 12/Week 12 notes.md

Document databases and MongoDB

This is a middleground between key-value DB's and RDB's.

• Documents structres:

  • less strict
  • nested
  • repeated
  • order-sensitive

• Less interlinking - When we treat the documents as a database, any link between documents is not relevant for retrieval, at first.

• ... or linking less importan for retrieval.

Document formats

• Markup languages for text
• Markup languages for other data
• Bespoke formats
• JSON (JavaScript Object Notation)

JSON differences

• Simple expresison of standar data structures

• Useful for:

  • Data persistence (serialization)
  • Data exchange

• Very similar to JavaScript:

• Data types are similar (number, string, decimal)

• We can't name variables and reuse them

• In JSON, keys require quotes, in JS it's optional

• No named structure/cross reference

• Object keys take quotes in JSON

MongoDB

• Document-oriented model
• Distributed (by sharding, horizontal partitions)
• Open-ish Source... (copyleft)
• Operates on objects e.g.

  movieData = [{title: "Ant-Man", year: 2015, actors: ["John", "Bob"]},
  {title: "..."}]
  

  It's not a prescribed structure, but it's ok with MongoDB

MongoDB: INSERT

db.movies.insert(movieData)

MongoDB: SELECT

db.movies.find({year: 2015});

db.movies.find({title: /*Man/});

db.movies.find({title: /*Man/, year:{$lt: 2010}}); less than 2010

MongoDB UPDATE

db.movies.update( {title: "Ant-Man"}, <- find {$set: {year: 2015}} <- update );

Or update many entries

db.movies.updateMany( {actors: "Elvis Presley"}, {"actors.$": "Elvos Proslee"} );

Add an actor:

db.movies.update( {title: "Ant-Man"}, {$push: {actors: "Michael John"}} );

MongoDB

• Fairly SQL-like
• Reduces joins in treelike data
• Integrates well into code
• Has suffered security and reliability issues