Errata

In this section we said that the simple aggregation functions use an optimized internal path that is faster than the iterators, and that the engine replaces a single-column SUMX with its non-X version. We actually had this backwards.

SUM is shorthand for SUMX, so SUM ( factorders[_Quantity] ) is internally translated into SUMX ( factorders, factorders[_Quantity] ), and there is no separate faster path for SUM.

The closing sentence of the section repeats the same reversed direction, and the first paragraph also has a typo (“expectation” should be “exception”).

To fix this, please replace the first paragraph with:
“Iterator functions can have a performance impact on large tables when the expression they evaluate for each row is complex. A simple aggregation function such as SUM is shorthand for its X version over a single column, so the two have the same performance. We recommend using the simple aggregation functions when you are aggregating a single column, since they are shorter and easier to read.”

Please also replace the closing sentence with:

“They both produce the same value with the same performance, since internally the first one is translated into the second one.”