How the DAX Optimizer service works

DAX Optimizer analyzes a VPAX file that contains information about a Tabular model:

  • Table and column names
  • Measure definitions
  • Relationships
  • Statistics about the number of rows in tables and the number of unique values in each column

However, it’s important to note that the VPAX file contains no actual data from the model. It only holds metadata about the model’s structure.

These are the typical steps of optimization:

  1. Create the VPAX file by using tools like DAX Studio, Tabular Editor 3, or Bravo for Power BI.
  2. Upload the VPAX file to the DAX Optimizer service. The file is stored as a version of a model in DAX Optimizer.
  3. Run the analysis of the model version. This operation consumes one “Run” from the contract associated with the model.
  4. Review the recommendations provided by DAX Optimizer:
    • Start from the issues with the highest relevance.
    • Mark as ignored the measures that cannot be optimized (false positives)
    • Mark as fixed the measures that you modify based on the recommendations.

Once you complete the review of the measures involved in the more important issues, you can restart the analysis from Step 1 to make sure that the measures that have been fixed no longer generate recommendations.