What is a dense measure? and why you should avoid it.

There’s not enough memory to complete this operation.

There’s not enough memory to complete this operation.

Examples of dense measures:

[measure] = 1 [measure] = COUNT('Sales'[ProductKey]) + 0 [measure] = If([Total Sales] <> 0, [Gross Margin] / [Total Sales], 0)

Power BI queries assume dimension modeling: dimension columns from different tables added to a visual form a cross join space in which measures are evaluated.

In other words, the presence of relationships to connect all the tables together is ignored when query space is formed.

A measure is sparse when its values are mostly blanks. It is dense when it has non-blank values in most of the cross join space.

For example, if you add product, customer, date columns and total sales measure to a visual. The measure is sparse as only a small fraction of combinations of product, customer, and date have sales, hence return non-blank values.

But a measure that returns constant 0 is dense because it returns non-blank value 0 in the entire query space of the cross join of the three columns.

- Jeffrey Wang