SUM vs LOOKUP performance considerations

Justice

Is there any performance cost difference between a) SUM on a list line item in your data module, and b) dimensioning your data module by the list, then LOOKUP on a dimension?

I am asking because I have a data module that is dimensioned by several large lists, creating sparsity. I can reduce sparsity by removing some of those lists from the dimensions, but would required changing subsequent references from lookups to SUM.

devrathahuja

@Justice - Hi, what is the setup exactly like? Since you mentioned, you can reduce it by removing some of those lists, which dimension is it that can be removed and wouldn't change the data in this case?

Additionally, you can create a test blank model & recreate the scenario with 5 list members & both the approaches simultaneously and look at the calc effort column. The % calcs would vary a lot since it will be a very small model but you'll find out what's the most optimal way in this case.