What I learned from this simple CRUD project and John Smilga lecture:

• collections to represent group of items (instead of tables)

• document-based which represents a single item (instead of row)

• use Mongo Atlas since ultimately gonna use it.

  • steps:
    1. build cluster
    2. create first db user
    3. add IP address (whitelist for all for simplicity)
    4. connect

• If deploying to heroku, IP address must be set to allow everywhere

• Database > Collection > Document

• MongoDB has dynamic schema, i.e. 2 documents in a collection don't necessarily share the same schema 😅 Therefore we should use Mongoose to create consistent schema.

• PUT VS PATCH

• PUT: replace the existing resource

• PATCH: partial update

• Don't set up one type of response or one route then go totally different in a different route, this can avoid confusion to user *

• Always set PORT to process.env.PORT || 3000 to ease up deployment