Begon damn dots (or how to remove nulls in an OLAP cube)

Shane Gibson — Wed, 01 Jul 2009 22:01:31 +0000

If you create a SAS OLAP cube in 9.1.x and their is some some sparsity in the values (when is there not) then you will end up with lots of .’s when no values exist for an intersection.

This is not really a problem till you use Web Report Studio to report off the cube and then the user exports the report to excel in a non formatted format.

When they do this the . is passed to excel and then Excel refuses to apply formulas nicely when creating them on these cells (funnily enough the Excel sum formula does work for these).

So in this case you want to force WRS to show .’s as 0 and then they export as 0′s and excel is happy (or you can force the users to manually search and replace them each time they export if you are so inclined).

So the work around for us was to create a custom measure in the information map that changed all the .’s to 0′s. the MDX for this was:

IIF(<>.Value = null , 0 , <>)

Of course the down side of this is that the OLAP server has to do the calc each time it renders the values rather than a look up.

In hindsight we could have created a custom measure in the OLAP cube that did this and it would be calculated during the cube build rather than during the report rendering.

Blogging about all things SAS » SAS Web report Studio exports null

Begon damn dots (or how to remove nulls in an OLAP cube)