In DAX (Data Analysis Expressions), calculations can be categorized into three types based on their aggregative behavior across different levels of a data model:

Additive Measures

These measures aggregate consistently across all levels of a dataset. They work well with standard aggregation functions like SUM, COUNT, or AVG.

Example: Total Sales

= SUM(Sales[Amount])

If you sum sales at the country level, region level, or individual transaction level, the total always remains consistent.

Works well with simple arithmetic operations (addition, multiplication).

Non-Additive Measures

These measures do not aggregate properly across different levels of a dataset. They usually involve calculations like DISTINCTCOUNT, MEDIAN, or PERCENTILE, which don’t sum up directly.

Example: Unique Customers

= DISTINCTCOUNT(Sales[CustomerID])

If a customer buys multiple times in different regions, counting them at a national level won’t be a simple sum of the regional counts.

Semi-Additive Measures

These measures aggregate correctly for some dimensions but not all. A common example is when values should be aggregated over some dimensions (like product categories) but not over time.

Example: Ending Inventory

= LASTNONBLANK(Sales[Date], SUM(Sales[Stock]))

Inventory counts are meaningful at a single date level but shouldn’t be summed over multiple dates.

Examples of semi-additive calculations:

LASTNONBLANK() (Last known value over time)

FIRSTNONBLANK() (First known value over time)

AVERAGEX() (Averages that depend on different granularities)

Comparison Summary

Would you like more real-world examples related to property management?