If I have a table such as...:

YYMM | Sales
1201 | 10
1202 | 20
1203 | 30
1204 | 40

... then I might describe this as having a time dimension. If I wanted to calculate a year-to-date value of Sales, would "year to date" be a different dimension to time, or is it actually the same and the additional measure implies the different meaning? Such as:

YYMM | Sales | SalesYTD
1201 | 10    | 10
1202 | 20    | 30
1203 | 30    | 60    
1204 | 40    | 100

Year to date sales is a different measure. Depending on your platform (OLAP systems, for example) it may be purely calculated on the fly. If you have an appropriate unit of analysis to do so, store it, otherwise calculate it on the fly.

  • May I ask what you mean by an appropriate unit of analysis? I think you mean something like "an appropriately grained dimension". Is that right? – jl6 Apr 25 '12 at 9:53
  • Combination of dimensions (e.g. product, sales agent ...) and a rollup of time. What you calculate must be additive across anything you want to slice it by, so you can't have multiple incremental YTD figures for the same thing appearing in the slice. – ConcernedOfTunbridgeWells Apr 25 '12 at 10:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.