Skip to main content
Aggregation combines several values into one result. In formulas, aggregation can happen in three related ways.

​
1. Property aggregation

A property has a default way to roll underlying values into a model cell. Metrics normally use sum. Dimensions normally use first because they describe a group rather than a numeric total. This behavior applies when source data contains several records for the same model position.

​
2. Multi-value references

A formula reference can deliberately select several values. For example:
This selects Revenue across all Department values. If no aggregation function is written, a multi-value metric lookup uses sum. It is clearer to write the intended aggregation when the choice matters:

​
3. Aggregation functions

Aggregation functions let you choose how several values become one result. Supported functions include:
  • sum
  • average
  • min and max
  • count
  • median
  • stdev and var
  • first and last

​
Example

If Department Revenue is Sales 50, Marketing 30, and Engineering 20:
  • sum(Revenue[Department: ANY]) returns 100
  • average(Revenue[Department: ANY]) returns about 33.33
  • max(Revenue[Department: ANY]) returns 50
  • count(Revenue[Department: ANY]) returns 3 when all three values are present

​
Write the business meaning

Use the aggregation that matches the question. Totals usually use sum. Rates and ratios often need to be calculated from their underlying totals rather than averaged after the fact.