Dimension Build Setting Tab:
1. Here you will specify which Essbase dimension or dimensions you are building with this rules file. Yes, you can build two dimensions with one rules file, this is more advanced, so we’ll leave it for another topic.
2. If the member is already existing, do you want to allow moves (new location in heirarchy), allow property changes (change the aggregation from ~ to a +), or allow formula changes. Other options will appear based on given dimensions selected, like UDA changes.
3. Build Method – this is where you specifiy whether or not a given dimension is to use Generation References, Level References or Parent/Child references. As I mention in the Field Properties section, using Parent/Child references when available is the recommended approach. You can also check the box to process null values (this is really applicable when building your dimension using Generation references)
4. Upon building the dimension, you can sort the members Ascending, Descending or no sort at all. Be careful of this option. If you accidentally clicked on a dimension name, say “Scenario”, but you were not building the Scenario dimension with this rules. Then if you checked the Sort Ascending button, it WILL sort your Scenario dimension even though you did not intend to.
5. This will “Merge” any existing members, plus the new ones coming in or “Remove unspecified” will delete any member that’s not included in the data source being loaded.
Dimension Definition Tab:
The key feature here is that you have the ability change a Hyperion Essbase Block Storage database’s dimension settings, but the one I’ve used the most in practice is changing the Dense/Sparse settings via a load rule.
- Right click on any dimension name (select “Edit Properties”)
- If you dimension is currently Dense, but you want it to be Sparse (say for calculation performance), then you change the “Configuration” setting to “Sparse”.
- Note: You can add implement this feature on a existing load rule (even if that load rule is building a different dimension) or create a seperate one, but your data source must contain some sort of data as it can not be blank.