If you are unable to create a new account, please email support@bspsoftware.com

 

News:

MetaManager - Administrative Tools for IBM Cognos
Pricing starting at $2,100
Download Now    Learn More

Main Menu

Power cubes and schema

Started by juffin, 27 May 2014 07:40:39 AM

Previous topic - Next topic

juffin


Hi

I have never worked on cubes, so please forgive me if this is silly..
In DMR and Relational models we follow star or snow flake schema
But I heard somewhere that power cubes are not using snow/star schemas.
I know it's dimensions,levels,attributes etc..but how how is connected to fact and other dimensions..
Where can we see or create the relationship..?

Appreciate any answers..Adv Thanks for your reply
Jf

bdbits

Power cubes are snapshots of data at the time they are created. Within Transformer, you provide data sources (queries really) that run at cube creation time. In my work we usually use Cognos packages but you can read a variety of data sources. In any case, the data from the queries is transformed into the facts and dimensions according to how you have modeled them. So the cube once built is not using any database schema whatsoever; it is self-contained.

If you want to see the particulars of how the cube is defined, you will need to open the Transformer model.

I would note that the data source schema will have a definite effect on cube build time. You may still want a warehouse underneath your cube model, rather than building it from a transactional application database.

cognostechie

Quote from: juffin on 27 May 2014 07:40:39 AM
Hi

I have never worked on cubes, so please forgive me if this is silly..
In DMR and Relational models we follow star or snow flake schema
But I heard somewhere that power cubes are not using snow/star schemas.
I know it's dimensions,levels,attributes etc..but how how is connected to fact and other dimensions..
Where can we see or create the relationship..?

Appreciate any answers..Adv Thanks for your reply
Jf

A Cube does not deal with relationship. It works with 'association'. The ideal way to model a cube is to provide separate queries for Dimension and Fact information . When the cube is built, it associates the Fact data to all levels of the dimensions based on
a common field name. Ex: Product Dim query provides Product Type, Product Line and Item-No. Fact query provides Item-No and measures. The data for the measures will be associated with the levels of the Product Dimension with the Item-No.

juffin