Understand Data Records
-
Any member not within the data record aggregation path will be bypassed. This means that a datasource can contain a complete set of records. BI Blend can then bypass those records by “filtering” or selecting an aggregation point that excludes those record sets.
-
Attribute Members can be aggregated by being associated as a base member in a UD8 hierarchy. If the source record Attribute does not find a base target UD8, no error message is presented, and the record is ignored. This allows the data set to be flexed easily to adapt to reporting changes. Users should review the messaging for by passed Aggregations. If the member is not within the aggregation path it will be ignored.
Use of Common Members
Common Members reference metadata designs which have the same member names across dimensions. An example of Common Members is where Dimensions such as UD2 and UD3 both have members called “Top” with children as “None”.
-
Caution when using common members (Top / None) across dimensions as the common members may cause inconsistent results.
-
If fully summarized intersections are required, the designer should consider selecting another Blend Unit.
-
Limiting a Blend Unit to a member may not be optimal for BI Blend processing but will yield fully aggregated dimension results.
-
If duplicate records are encountered, accept the duplicates but use aggregation queries when you consume or query the BI Blend Table (Group By on dimensions while performing a sum on Measures).
-
Unique top members across dimensions are preferred, change dimension aggregation information to pick a parent that does not include common members, such as None or Top.
