If you have activated the collection option for query runtime statistical data, every navigational step that requests data from the database is saved. This includes the characteristics, navigation attributes and hierarchies that are involved. First the system saves this statistical data in database tables. You can then load this data into BI statistics for evaluation purposes.
You can also use this statistical data for optimizing aggregates. The advantage of having proposals from BI statistics as opposed to proposals from BI queries is that the actual user capacity is taken into account here.
We recommend that you use this function if representative statistical data already exists. You run this function at regular intervals to modify the aggregates in accordance with changes to user actions.
You can evaluate the data saved in BI Statistics (database) or BI Statistics (InfoCube) by selecting the menu path Proposals → Proposals from BI Statistics (Tables) orProposals from BI Statistics (InfoCube). You can restrict the analysis to a subset of the data by specifying an interval for the start time or runtime of the query.
After the data has been read from BI Statistics, the optimal aggregate is determined for every navigation step, and a list of the different aggregates is created.
For aggregates in a component that vary only in terms of their selection type, all aggregates with the selection type hierarchy (‘H’) or fixed value (‘F’) are replaced with an aggregate that is grouped by characteristic values (‘*’). This is possible as long as the aggregate has already been proposed.
The list of proposed aggregates has the same structure as the proposals from BI queries. However, the list is generally longer.
You can modify or delete the proposed aggregates.
Optimizing Proposed Aggregates