Is it possible to create a combined list by selecting members from two different tables and columns?
Is it possible to create a combined list by selecting members from two different tables and columns?
In formula creation, I'm getting the following message
"Sets must use members from the same table and column."
12 replies
-
Hi
A list is always designated to a specific table and column /hierarchy.
Can you perhaps explain what you are trying to achieve? I'm confident we can achieve the required functionality.
-
Example : SELECT A.number1, B.description, C.description FROM TableA A JOIN TableB B ON A.number1 = B.number1 JOIN TableC C ON A.number1 = C.number1;.
I'd like to load only the descriptions in the list that match from both tables based on the values from Table A
-
Can explain what your model looks like and what you're trying to do? Do you want one list that will have all the values from B.Description and C.Description? what will you do with it consequently?
You could create a list of all B.Description and then filter exists on all C.Description. So you'll get only ones that exist in both dimension - but I sense I don't have the actual use case.
Can you share on your model structure and the functionality you need? -
It is important to remember that the Pyramid Semantic models that represent your underlying relational data are fully dimensional in nature, i.e., each attribute is treated independently of all the others. Thus in Pyramid Lists, the members of the list are referring to the list of unique values that make up that attribute. There is no concept of the JOIN construct you are referring to when manipulating Lists in Pyramid.
When using Pyramid semantic models, then Lists can only contain elements from the same Column / Table or, in Pyramid terms, Hierarchy / Dimension. This how the Pyramid expression language, PQL has been designed. Note however, that if you are working with SQLServer Analysis Services semantic models, using MDX, this does support Sets from multiple hierarchies.
As Moshe says, we'd need to know a bit more about your model and final output before we can help further.
Hope that helps.
Ian