More on DW and SSAS Processing

 This page contains:

• Processing Modules/DW and SSAS multidimensional databases
• Typical SSAS error messages

 For Tabular SSAS modules check the tutorial for Tabular SSAS.

Processing Modules/DW and SSAS databases

Profitbase Studio Modules is organized according to the SSAS model. 

• Module Group associates to the SSAS database
  • Dimensions are shared in the Module Group and in the SSAS database.
  • All tables/views with columns (the DW schema) are registered in the SSAS database.
• Module Sub Group associates to SSAS Cube.
  • All SSAS BI are defined in Calculation Scripts in a Cube. PS defines this in the Module Definition UI but it will be configured in the Cube.
  • SSAS Actions are defined per Cube.
  • Dimensions used in SSAS Measure Groups will also be defined in the Cube.
• Module Definition associates to SSAS Measure Group.
  • Fact tables and dimensions are associated to the Measure Group through Measures and Dimension Usage.
  • Partitioning and Direct Queries can be associated with the Measure Group to optimize resources and aggregations.

 

 There is two main reasons for processing Modules in DW and SSAS databases:

• Configuration changes; this is needed while creating or changing configurations like adding Module parts, changing Names, Translations, Business Logic, Measures, Dimension Usage etc.  
• Data changes; this is the daily updates of data from Data Sources and Dimensions that needs to be updated in the Modules DW and processed into the SSAS database.

The Modules processing is organized in 2 parts:

1. Preparing the Module in the Data Warehouse database (dw). This creates a unified set of tables and views in dw. Whether it is a table or view, can be configured. Dimensions that have reduce scans, needs to be a table.
2. Adding configuration and processing of SSAS databases.

Typical SSAS error messages

"A duplicate attribute key has been found when processing.." or
"The attribute does not have any members.." or 
"The attribute key can not be found.." 

1. Check the fact data source and the dimension tables. Is the data correct ?
2. Turn on Member Scan if it is the key attribute that has the error.
3. If it is a column linked to a time dimension ensure that the fact source column only contains the date part (not time). Use Date data type or use FLOOR(CAST( AS float)). Also ensure that the span of the time dimension covers he data range.
4. In Wide dimensions, check the attribute relation cardinality - change from one to many.
5. In Wide dimensions, set the attributes 'Value If Null' on the key column (to e.g. *NULL).
6. In Wide dimensions, add extra key column to the attribute to make it unique.
7. If the value reported is *NULL or *EMPTY, add a row with this item key to the dimensions modification table. 
8. Add Advanced SSAS Settings to the dimension to configure error handling.
9. Add Advanced SSAS Settings to the processing command to configure error handling.