Recommendations for the Definition of Queries 

The way in which queries are defined directly influences performance during data analysis.

The following recommendations help you to take performance aspects into consideration when defining queries.

●      If you have large volumes of data, use aggregates or the SAP NetWeaver BI Accelerator.

●      If you use DataStore objects and InfoSets in queries, the performance is not as good as if you use InfoCubes.

●      The response times of queries are shorter if they display strongly aggregated values that can then be drilled down further using free characteristics. You should therefore use as few characteristics as possible in the rows and columns.

●      Define calculated and restricted key figures in the query definition rather than locally in the executed query.

●      Keep the expected query result as small as possible (maximum of 1000 rows).

●      In Web applications, using graphics affects the runtime of the query.

●      If possible, execute all calculations that are to be executed before the aggregation (such as the currency translation) when the data is being loaded.

●      Suppress results rows if you do not need them.

●      Restrict time characteristics ideally to the current value.

●      Use the caching options. Use the OLAP cache for BEx applications. Use the MDX cache to precalculate BEx queries in the background, which then serve as data providers for other applications.