Queries extract a defined set of records. As long as no filter has been set, the records are queried without restrictions. This means that they contain all rows of the analysis realm defined. Whenever you set a filter, certain rows will be filtered out. The number of records is reduced according to the filter you set.
A filter is always applied to one criterion. The filter specifies your preferred and suitable criterion values. If a filter is set only those table rows with matching criterion values will be selected. The filter is then transferred to other tables, in which only those rows are selected that are linked to at least one selected row of the source table. The filter is transferred recursively to all tables relevant to the query.
Example
In this example, filtering for an administrator by the name of Kunz returns only the values of those rows that are linked with the corresponding row in the Administrators table.
Query of the application and the leasing rate for each application filtered by the name of the administrator Kunz
Rows not linked to a row selected will not be included in the record selected. Therefore, the row of the Cognos application will not be included in the query result as long as the query contains more than just the Application table criteria. If the query used the Application criterion for iteration and aggregated the licenses, Cognos/30 would be included in the query result.
Query result with an additional filter set to Munich or Berlin
The filters supported and their semantics in data analysis are similar to those in PPM Process analysis. For example, text or period filters are available, and you can set filters for measures. You can combine multiple filters and use result filters for iterations, etc.