Introduction to NoSQL

Jake Roach

AI Engineer

Traditional relational data stores (RDBMS)

Organize data in tables, using columns and rows
Leverage SQL to manage and query data
Enforce integrity through constraints on databases and tables

Image containing MS SQL Server, MySQL, and PostgreSQL logos.

What is NoSQL?

Definition: NoSQL stands for "not only SQL", and is a set of data storage tools and techniques that allows for structured, semi-structured, and unstructured data to be stored and retrieved.

Characteristics:

Allows for wide variety of data to be stored and retrieved
Less rigid schema
Better scaling and performance

NoSQL data stores

Tabular

Books table in a Snowflake database.

"Rectangular"
Using columns and rows

Non-tabular

{
  "title": "Python for Data Analysis",
  "price": 53.99,
  ...
}

"weather": "sunny"

Semi-structured format
More flexible schema

NoSQL data stores

Column-oriented databases

A NoSQL data store that stores data by column, rather than row, and can be queried with SQL-like syntax. Allows for faster querying of data, especially when running analytical queries.

Use case: big data, analytics workflows

Snowflake logo.

Document databases

NoSQL data storage tool used to store semi-structured "documents"

JSON format
Less rigid schema

Use case: user-generated data (reviews) and real-time analytics

Postgres JSON logo.

More NoSQL data stores

Key-value

A NoSQL data storage tool that stores data as a collection of key-value pairs:

Simple data that is written and read at a high frequency

Use cases: IoT (Internet of Things) data, mobile applications

Redis logo.

Graph

A NoSQL data store that persists data in a network of nodes and edges.

Nodes represent entities
Edges represent relationships between entities

Use cases: social networks

Let's practice!

Introduction to NoSQL