SUM vs SUMX in DAX: Why a Row-by-Row Calculation Needs an Iterator

By Michael Nocito, data analyst · Published August 10, 2026

By the end of this page you can look at any DAX aggregation and say which of the two it needs, before you write it. You will also be able to spot the wrong one in somebody else's model, including the version that returns a number close enough to pass. It takes about twenty minutes. Every figure below is printed next to the arithmetic that produced it, so you can check any of them on paper.

Here is the move. If the number you want to add up already sits in a column, use SUM(Orders[Units]). If it has to be worked out on each row first, use SUMX(Orders, Orders[Units] * Orders[UnitPrice]): a table, then the calculation to run on every row of it.

The short version: SUM adds a column that already exists. SUMX builds a value on each row, then adds those up.

The mistake that follows from missing this is worth seeing as a picture, because the arithmetic explains itself once you have.