Aggregates are subsets of InfoCube data, where the data is aggregated and kept in a structure of InfoCube, which the purpose is to reduce the amount of data that must be read in the database and accelerate query response time.
This blog gives some useful tips on aggregate creation and maintenance for tuning query performance.
1.MASTER DATA MAINTENANCE BEFORE AGGREGATE CREATION:
To reduce the amount of data the aggregates can be filtered on characteristic level with fixed value or hierarchy. To be able to select a fix value, the needed one should be already maintained in master data table for that object. Therefore, Necessary master data maintenance should be taken into consideration before the creation of aggregates.
2.EXPERT MODE FOR AGGREGATE MAINTENANCE:
Aggregate Maintenance screen enables developers to include the characteristics included in dimensions of the infocube. In case of an aggregate need which includes a navigational attribute the expert mode should be switched on. To be sure that there is no missing objects for the aggregate required, always working in expert mode on is a better approach. To Switch it on go to Extras->Switch Expert Mode On/Off.
3.MASS ACTIVATION OF AGGREGATES:
After the creation of aggregates, they should be activated and filled. It can be done with the Activate/Fill Button in aggregate maintenance screen. However if there is more than one aggregate at the moment to be activated and filled, this button means transport requests for each aggregate. And in case of many aggregates for an infocube, it may cause loss of time. A program enables mass activation and filling of aggregates for the selected infocube without dealing with import requests: SAP_AGGREGATES_ACTIVATE_FILL. This program takes infocube name as an input and three options: 1) Fill Agregates 2) Display Warnings 3) Display Errors. Second and Third Option basicly enables the warning and errors messages to be seen during Aggregate activation. First Option enables filling process of aggregates after activation process. Executing this program in background reduces loss of time for this activation process. After the program execution is completed, the transport can be handled via Transport Connection tab of Datawarehousing workbench in case of transport need.
There may be such cases that all aggregates should be inactivated. There is also a program for this purpose: SAP_AGGREGATES_DEACTIVATE. This program deactivates all aggregates of the selected infocube. There are also two options to display error and warning messages.
It is possible to copy aggregates to a cube with program RSDDK_AGGRCOMP_COPY. Here the structure difference of cubes should be taken into consideration. The common structures are only copied to the other cube, therefore if there is some more infoobjects needed in the destination cube, they should be added manually.
5.INITIAL FILL FOR AGGREGATES:
Whenever an infocube’s data is loaded, a roll-up of aggregates can be set automatically. However, the first fill of aggregates should be handled manually. For this purpose aggregate maintenance screen can be used but there is an easier way to do it. There is a process type that is called “Initial Fill of New Aggregates”. Creating a process chain for it and executing it will make life easier:
6.ATTRIBUTE CHANGE RUN:
Whenever a new upload of characteristic master data occurs, aggregates that includes the infoobject should be realigned by Attribute Change Run. But this realignment is a heavy process that should be taken into account. Change run should be scheduled carefully.