Databases

Mongo DB: Using compass for "aggregate" querying

• Mongo has a system of pipeline which allows to chain several filters and transformation operations
• Compass has a system that generates this pipeline for different languages, like java
• run your aggregate query in compass and realize a java, node or python integration of it-

Implementing a model with MongoDB

• Documents require no schema in MongoDB
• It means the database structuration is totally free!
• Collections can contain really various kind of data

Implementing a model with MongoDB(2)

In MongoDB it is possible to add successively those 2 documents:

db.users.insertOne({
    name:'Estelle'
})
db.users.insertOne({
    email:'miguel@somecompany.com'
})
db.users.find();
    

which leads to this following result

Implementing a model with MongoDB (3)

Is it possible to use this kind of data in an application?

Answer? No!

• It is mandatory to have minimal data filled for each record
• The UI requires a certain structure, so data need to reflect at some point that structure
• It is better to see the schema-less aspect of mongoDB as a flexibility offered by the system

Schema-less DB : a viable approach (with a bit of Schema!)

What's a good approach to converge to a usable result?

• have a "fixed" structure, which represents the essence of the entity
• have a "variable" structure which can be optional

Modeling exercise (2)

Nested documents or references?

• No predefined rules, relies on the application usecases
• Depends mainly on the access frequency to the data
• one to one relationships can surely be nested always
• one to many relationships can be implemented with embedded, references or subset strategies

Modeling exercise (3)

• Define the different cases of the application
• Define the collections
• write the mongo queries to achieve insert/search queries