You could create a standard hierarchy, using a second level to aggregate the
base values, or you can evaluate a more dynamic approach, as explained by Tom
Chester's article on "Banding Dimensions" (www.tomchester.net).
You can also create a one-level dimension with the band values and source
the cube with a fact table view that calculates the correct band for each
record.
Each solution has pros and cons, depends on your exact scenario.
HTH,
Brian
www.geocities.com/brianaltmann/olap.html
"toms" wrote:
Quote:
Hello.
I have one level dimension containing data from 0 to 30. There is about 40
of different values. I want to discretize it in a sections like 0 - 5; 5-10;
10-15...... etc. How do this ? |