Both Cube and Information Sources data sources are implemented to provide the end user the same user experience. But underlying implementation uses different technologies.
While Information Sources uses SQL (Structured Query Language), Cubes uses MDX (Multidimensional Expressions).
Due to this very different nature of technologies and the vision to provide the same user experience with both data source types there are certain limitations.
This page describes such limitations with each data source type.
Information Source should have at least one measure item to access within Lobby
Only AND conditions are supported between filters
When ALL check box is selected, the filter will be removed.
No ordering will be made unless explicitly defined.
Data will be aggregated using summation for Measure Items (Oracle SUM function) and will be grouped by (Oracle GROUP BY clause)
Dimension Items and Light Items used in the query.
LoVs will display data as string values after formatting using the default formats for the database session. This might result in inconsistencies on how data is displayed. For an example, date values will be formatted using the client OS regional settings when executed while the values in the slicers will be formatted using the database NLS settings defined for the database session.
Only AND conditions are supported between filters
If there are dimension items, levels with the same id in same SQL Server Analysis Services server, it is assumed that these dimensions shares same data across. i.e. The keys will be same and the filters will be applied across the Cubes when sliced from Lobby pages
The mapped users should have access to the Cube/Data to fetch data. Unlike for Information sources, this is the case for both design time and run time.
Only on-premise SSAS installations on the same domain are currently supported.
When ALL check box is selected, the filter will be removed from the query.
Sort operations on Dimension Items will be always based on String value
MEMBER_CAPTION. Even though the underlying value is a number, sorting
will always be performed on the MEMBER_CAPTION (which is always a
String)
Sort Operations can be performed only on the Lowest level of Multi Level Dimension Items
Data will be grouped according to the definition of the Measure Item in the cube. Grouping by the Dimension Items will be handled in SQL Server Analysis Services.
LoVs will display data as string values after formatting using the formats defined in the Cube. This might result in inconsistencies on how data is displayed. For an example, date values will be formatted using the client OS regional settings when executed while the values in the slicers will be formatted using the formats defined in the Cube.