Considerations for Changing Cube Consolidation Algorithm Type
Good afternoon,
Can anyone share their experience of the potential impacts of switching the Cube Consolidation Algorithm Type from "Standard (Calc-on-the-fly share and hierarchy elimination)" to "Org-By-Period Elimination"? I am balancing the decision between making a second cube with the different algorithm type vs changing the algorithm type of our current primary cube.
I'm curious if anyone has any experience with the algorithm type outside of OneStream included documentation (see below) like consolidation time performance impacts.
"Org-By-Period Elimination
The Org-By-Period Elimination Consolidation Algorithm Type uses the calc-on-the-fly Share – as in Standard – but has unique elimination considerations. When determining if the data cell’s IC Member is a descendant of the Entity being consolidated, this Consolidation Algorithm Type considers the position of the Entity in the hierarchy and also checks the Percent Consolidation for every relationship down the hierarchy. If Percent Consolidation is zero for the particular relationship, the IC Member is determined not to be a descendant of the Entity.
In comparison, the standard elimination (hierarchy elimination) only considers the position of the Member in the Entity Dimension hierarchy. Standard elimination is the default approach and does not consider Percent Consolidation."
Some more detail of why my org needs to switch: my organization has added an alternative hierarchy where previously 100% consolidation, full consolidation type, entities are now being reported as 0% consolidation, equity consolidation type. The system is still running intercompany eliminations on the equity entity when we don't want it to. I want the current hierarchy eliminations to remain as is, but need the alternative hierarchy to not run eliminations on equity. 0% con entity. I have tested the new setting out in a test environment and haven't found any issues, but this seems like a pretty big change to be making to our system without exploring more.