Correct. Get the distinct customers with their salary, add a dummy column called “flag”. Let flag=1.
Union this pre-aggregated data with the original table. In the original table, the dummy column “flag”=0.
After then, the calculated field in above post should work.
If you visualize the data with KPI, or a table without any dimension, you can avoid distinctcountover with avgif(salary, flag=1). If you visualize the data with dimensions in the visual, the distinctcountover can’t be avoided. Enjoy it! In near future, we might have a better solution. We will contact you again after the new feature release.