- Calculated columns should be used carefully because they increase model size and memory usage.
- In one project, adding multiple calculated columns on a large fact table slowed report refresh.
- Each row stores the calculated value, unlike measures that compute on demand.
- High-cardinality columns like transaction IDs consume significant memory.
- They also increase storage in Power BI Service, affecting performance.
- Using measures instead of calculated columns reduces data duplication.
- If a column can be computed in Power Query or at source, it is more efficient.
- So unnecessary calculated columns can degrade both refresh and query performance.
Why should calculated columns be used carefully?
Updated on February 25, 2026
< 1 min read
