Relational or MDX

[dsds]

I have a requirement where i have some KPIs against Region. We have mulitpl mockups as below

1st mockup
Country Region  State KPI1 KPI2


2nd mockup

Region State KPI1 KPI2


3rd mockup

State KPI1 KPI2

The data would be available at the State level and we are asked to aggregate it up. Should i go by cubes or relational reports. My hunch is relational reports

You would have used a cube if the same column would have been used for each level of a hierarchy. But, here they are kept side by side which seems a relational fit. Secondly, there are no requirements that ask for complex dimensional functions. Is it good?

[blom0344]

Definitely relational by the look of it.  Of course by starting relational you still have the choice to add a dimensional view by means of DMR

[dsds]

The only reason i thought of MDX was since its moving up and down in the region hierarchy and to improve performance. But, on second thoughts its not a good enough reason. Since, its more of a drill through then a drill down.

[MFGF]

If you use Blom's suggestion of relational DMR, you will have the same MDX functionality available in your report as you would if you built a cube.

Cheers!

MF.

[dsds]

my question is more towards relational vs mdx given the report requirements. I am ok with creating a cube or DMR, but, only if the report requirements warrant it, and not if its not required.

[blom0344]

my question is more towards relational vs mdx given the report requirements. I am ok with creating a cube or DMR, but, only if the report requirements warrant it, and not if its not required.

Surely you do not build a model to build just the one (1)  report?

[dsds]

I quoted the type, but, i would have 20-30 odd similar reports