Which measure is most likely to perform best when filtering by a specific calendar year?

• A. CALCULATE( [Sales], Calendar[Year] = 2023 )
• B. CALCULATE( [Sales], FILTER( Sales, Sales[Year] = 2023) )
• C. CALCULATE( [Sales], FILTER( Sales, YEAR( Sales[Date]) = 2023) )
• D. CALCULATE( [Sales], Sales[Year] = 2023) )

Answer: (A)

Filtering by the calendar dimension’s Year column inside CALCULATE targets the correct time context using the established date table. This approach relies on the existing relationship between the sales data and the calendar, so applying Year = 2023 restricts the data to every day in that calendar year and then propagates that filter to Sales. It’s precise, scalable, and works consistently with other time-based filters or slicers that use the same date dimension.

Filtering the Sales table directly (whether by a Year column on Sales or by extracting the year from the date) bypasses the calendar dimension. That can work, but it couples the measure to a possibly redundant or less robust year attribute and may behave differently if you switch calendars or adjust date grain. Using the YEAR function inside a FILTER is also less efficient, because it evaluates a row-by-row expression on potentially large tables, rather than leveraging the optimized relationship with the calendar.

So, the calendar-based filter integrates cleanly with the model’s time intelligence, delivering correct, performant results across contexts.