LLearningPath
Data Modeling with Databricks SQL

5% domain · module 1

Star vs snowflake, grain, and SCD

Starts at 1:16:19 in this video — Dimensional modeling + SCD in the SQL Warehouse course

Star schema

A fact (measures at a declared grain, e.g. one row per order line) plus denormalized dimensions (date, customer, product). Analysts join on keys. This is the default BI model.

Snowflake

Dimensions are normalized into more tables (product → category). More joins, less duplication. Use when dimensions are large and shared.

Grain and SCD

Write the grain in a comment. Type 1 overwrites attributes; Type 2 adds a row with effective dates. “Customer segment as of order date” needs Type 2 (or a snapshot), not the current customer row.

The 3h course walks star schema and Type 2 SCD in Databricks SQL — watch from 1:16.

SQL modeling

Next module →