Skip to main content

​
Basic metrics (BASIC)

A basic metric binds a business name to an existing measure in a semantic model. For example, a model measure [Measures].[Sales] can be exposed as Sales with time, organization, or product filters. Publishing checks that the measure exists and the model version is available.

​
Derived metrics (DERIVE)

A derived metric calculates a value from one or more measures, such as profit margin. In an XMLA model, the formula is a value expression rather than a complete query: ([Measures].[Profit] / [Measures].[Sales]) * 100 The expression must reference a measure in the published cube, and the system checks its length and statement keywords. For complex SQL calculations in a relational model, add a calculated member to the cube first and bind the metric to that measure.

​
Which type should I use?

  • Use a basic metric when you need a common definition, unit, or fixed filter for an existing measure.
  • Use a derived metric, or create a calculated member in the model first, when you need a new value calculated from multiple measures.
  • Cross-model joins are not provided by the current metric query function. Split the models or shape the data at the source.
Click Query before publishing to verify the value, language, and result.