Context transition for aggregator instead of using existing row context

An aggregator function (like SUM, MIN, MAX, …) is executed within a context transition instead of using a column reference in the existing row context.

Remarks

The row context of an iterator is transformed into a filter context with the context transition produced by CALCULATE, providing a filter context that has only one row when the iterated table has unique rows. In this condition, the aggregation function only groups a single row, returning the same result that would have been obtained by using a column reference on the existing row context.

From a performance standpoint, the context transition consumes CPU and memory, so removing the need of a context transition and removing the aggregation improves performance. The optimization is possible only when the iterated table has unique rows, otherwise the result could be different.

Example

Remove CALCULATE and SUM so that the column reference is evaluated within the row context produced by the iteration over Store.

Original code

SUMX ( 
    Store, 
    CALCULATE ( SUM ( Store[Square Meters] ) ) 
)

Possible optimization

SUMX ( 
    Store, 
    Store[Square Meters] 
)