Thanks for the reply . Ok both are multi dimensional structures . then my question is why to go for cubes then ?. only impromptu report is enough then .Please clarify this doubt .
thanks .
I main reason for cubes is performance. Inpromptu reports query the source data and the aggregate the data that is reurned. In some cases where the dataset is smaller it might not be an issue but on large data sets impromptu reports can be slow. Impropmtu reports also lack an easy slice and dice functionality.
Cubes are self contained (exception being ROLAP cubes). When the cube is built all the data is preaggregated allowing for much faster query times. Also in the cube structure the ability to slice, dice and drill down is greatly enhanced.
Impromptu is a reporting tool, like the report builder in Power Play. Power Play Transformer, however, is a tool which constructs cubes. If Cognos had the mindset, they could enhance Impromptu to report against cubes as well as against relational databases through catalogs. Instead, they have a separate reporter in Power Play. In addition to aggregating data as MDX'er pointed out, cubes are very efficient in the handling of empty cells. A good example of empty cells would be at the intersection of hospital procedure and patient gender. At the intersection of these dimensions, you would not have any male pregnancies nor female prostage sugeries. Another example of empty cells might be sales of bagged ice by zip code in Alaska in the winter. Lots of nothing going on there !!!!
-------------------------
The trouble with doing something right the first time is that noboby appreciates how difficult it was.
- Steven Wright
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.