- Cardinality refers to the uniqueness of values in a column, like Customer ID vs Country.
- High-cardinality columns have many unique values, while low-cardinality columns have few.
- In one project, using Customer ID in a large fact table increased memory usage and slowed DAX calculations.
- High-cardinality columns also increase compression size in the VertiPaq engine.
- Relationships on high-cardinality columns can slow filtering and aggregations.
- Low-cardinality columns compress better and improve query performance.
- We optimized by avoiding unnecessary high-cardinality columns in visuals and measures.
- So cardinality affects memory, refresh, and report responsiveness.
What is cardinality and its impact on model performance?
Updated on February 25, 2026
< 1 min read
