From Chapter 2 - REST API, you probably noticed that data is not persisted yet 😢

Everytime we change code, server restarts (via nodemon) and our data is gone 👻

This chapter, we'll move our JSON files to the database so all the CRUD changes from our REST API are persisted.

First things first, you'll need a connection string from MongoDB Atlas. Follow this guide first if you haven't

MongoDB Atlas Signup

SQL vs NoSQL

SQL databases:

database > tables > rows

(database) data_db
(table) products

           id     name            brand      price
(rows)      1     toilet paper       X       199.99
            2     hand sanitizer     Y       299.99

NoSql:

database > collections > documents

(database) data_db > 
  (collection) products: [  
    (documents)         { "id": 1, "name": "toilet paper", "brand": "X", "price": 199.99, "type": "toiletries" },
                        { "id": 2, "name":"hand sanitizer", "color": "blue", "brand":"Y", "price":299.99 },...
               ]

But why noSQL?

• Dynamic schema for unstructured data (vs a defined schema in SQL)
• Add fields as you go
• Document-oriented, instead of relational
• JSON-like: represents objects closer than relational databases
  • e.g. they share the same structure with JS objects / JSON
• Speed for reading - basically a big hash with O(1) lookup
• Denormalization - e.g. get user, user.friends, user.friends.photos without joins
• When not:
  • When you need fancy queries (e.g. get all users with >10 friends)
  • Not good for ACID (atomicity, consistency, isolation, durability) compliant purposes
  • When you need more writes than reads

For details on the difference, see chapter footnotes