Chapter 17. Understanding a Relational Data Model

Relational databases have existed since 1970 and have introduced a lot of new concepts: tables, relationships, constraints, normalization, etc. The concept, and its implementation by various vendors (e.g., SQL Server, Azure SQL DB, or Azure SQL Managed Instance by Microsoft), is still successful and allows for a variety of use cases. That’s why you find both application databases (OLTP) and analytical databases (OLAP) implemented as relational databases.

This chapter guides you on how relational databases are different from Power BI and Analysis Services tabular. Due to these differences, you will learn that a relational data warehouse is the perfect addition to your analytical infrastructure in an enterprise environment. I explore techniques, use cases, and how to implement them in a relational database (managed and updated by SQL) to make the experience in Power BI and Analysis Services tabular great.

I introduce some basic concepts: that a data model consists of tables, that columns of a table can have different purposes (key or attribute), and how you can combine information that is spread out into different tables.

Data Model

A data model implemented in Power BI/Analysis Services tabular and one implemented in a relational database have many things in common. You store both data and metadata in it. The data is hosted in tables. Metadata explains how the tables form the data model. Let’s introduce the basic parts of a relational ...

Get Data Modeling with Microsoft Power BI now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.