Comments on: Precedence Project Part Two: VALUES()

By: powerpivotpro

powerpivotpro — Thu, 16 Jun 2011 13:24:04 +0000

EXCELLENT catch Derek. I’m pretty sure that’s because I filtered out blank years from the pivot. I will revisit when I get to my workstation

By: Derek C

Derek C — Thu, 16 Jun 2011 13:16:43 +0000

It’s interesting the two year values in the blue circle aren’t actually equal. $12,352,870$12,352,589 I know it is a miniscule amount in context, but I run into issues like this quite often. I spend a lot of time making sure my numbers tie when using DAX and continually run into little anomalies.

Does this imply that there is a dollar value of $281 in the dataset without an associated Periods[FiscalQuarter] value, but that does have a Periods[Year] value? Alternatively is there a filter the VALUES() function recognizes that the ALL() function overrides?

By: The Great Precedence Project (Filter Rules for DAX) – Evaluation Context in PowerPivot « George's BI Blog

Thu, 16 Jun 2011 06:44:50 +0000

[...] and Here, PowerPivotPro.com published an excellent blog about the DAX [...]

By: George Qiao

George Qiao — Thu, 16 Jun 2011 06:33:33 +0000

Actually, [Measure](ALL(Dim),VALUES(Dim[col])) is a shot from of Calculate([Measure],ALL(Dim),VALUES(Dim[col])), and a more generic form would be Calculate([Measure], exp1, exp2 ,… , expN).

The fact is that to calculate the [Measure], all expressions need be satisfied, which means exp1 and exp2 and….and expN.

In addition, any evaluation context on the slicer (as long as connected), report filters and rows/columns on the Pivot table will be considered as well, which means (exp1 and exp2 and….and expN) AND (Existing evaluation context).

I believe this is the underneath rule to explain all the above behaviour. But it is fantastic to see someone list the actual examples. It is hard to visualise the theory in brain and always good to have a quick sample to validate your thoughts. Thank you, Rob.