1 / 10
Use the arrows, swipe, or press ← → to move through the slides.
Snowflake vs Star Schema — when to normalise dimensions
Normalised dimensions reduce redundancy at the cost of joins. A side-by-side comparison with the cases where each wins.
The trade-off I'd flag
On a columnar lakehouse, storage is cheap and joins are not. Star wins by default; snowflake only when a dimension is genuinely shared and genuinely large.
