Why should high-cardinality columns be handled carefully?

< 1 min read

High-cardinality columns should be handled carefully because they consume more memory and slow down calculations.
In one project, including Customer ID in visuals increased model size and refresh time.
High-cardinality columns compress poorly in VertiPaq, affecting storage efficiency.
DAX measures using these columns, like DISTINCTCOUNT, are computationally expensive.
Relationships on high-cardinality columns increase filter propagation cost.
They also increase query execution time in both Import and DirectQuery modes.
We optimized by removing unnecessary high-cardinality columns or pre-aggregating them.
Careful handling ensures faster, lighter, and more responsive dashboards.

Power BI Essentials