Hi specialist,
when loading our customer-hieararchy from the R/3-system to the bw, automaticaly a root-node (in our case "standard-hierarchy" is added at level 0 of the hierarchy and the hierarchy itself is subordinated to that root-node.
In BW-reports this root-node often is very distrubing, because all more detailed characteristics in the drill down are dispayed double then. Once below the original top-knodes from the r/3 system - and again directly below the new, artificial root-node of the hierarchy.
So the idea was to remove that top-node from the hierarchy for getting better reports. That is possible by removing the top-node at data-loading-time using a customer-exit in the r/3 system.
The hatch is: When you try to search a single customer (in the BEx Analyzer) in that modified (root-node-less) hierarchy, the search takes much longer than it did before (with the "standard-hiearchy" top-node). It takes aprox 4 times longer to find a value and consumes about 1 GByte of roll-memory.
Did you ever meet that or a similar problem?
Any idea how to solute that issue?