Data modeling strategies in NoSQL databases

Proposed session for SQLBits 2026

TL; DR

Having a NoSQL database doesn't eliminate the need for data modeling, as it impacts scalability, performance, and cost. We’ll review JSON specs, create a sample data model, discuss NULL/undefined values, and explore partitioning and materialized views in NoSQL databases.

Session Details

Having a NoSQL database doesn't mean data modeling is unnecessary because of the no-schema concept. A well-designed data model is crucial for NoSQL databases, as it determines scalability, performance, and cost.
We will start the session by reviewing the JSON format specifications. Then, we will collaborate to create a sample NoSQL data model and understand how to manage NULL and undefined values in NoSQL databases.
In distributed databases, an important task is to select a property that evenly distributes data and traffic. You will learn the consequences of choosing a suitable or unsuitable partition key for your data models. We will conclude the session by creating materialized views that may assist you with cross-partition queries.
If you are working or planning to work on a project with a NoSQL database, this session is for you!

3 things you'll get out of this session

Data Type problems in JSON Why is the partition key essential in distributed databases? Guidelines for creating an effective data model for NoSQL databases.