The Local Cube Information Center
The Local Cube    Information Center    
www.cubeslice.com      

Differences between CubeSlice Relational ASSL and CubeSlice OLAP ASSL

CubeSlice uses two forms of ASSL (Analysis Services Scripting Language) to create local cubes. The two forms of ASSL created with CubeSlice are very similar. Here are the main differences:

1. CubeSlice Relational ASSL builds local cubes directly from the data in the source relational database, while CubeSlice OLAP ASSL builds local cubes from the Analysis Server cube. When choosing which data source to use, you have to consider

  • Permissions to access the Analysis Server and the source relational server.
  • The additional usage of the source server, either relational or Analysis Server.
  • The concurrency of the data. When using Relational ASSL the current data in the relational database is loaded into the local cubes. When using OLAP ASSL, data is loaded that is current as of the last time the Analysis Server cube was processed.
  • CubeSlice Relational ASSL is fully supported only when the source relational database is Microsoft SQL Server. When you are using other relational data sources some features, such as member slicing, are not supported with Relational ASSL.

2. When creating local cubes with CubeSlice Relational ASSL you have the option of removing unused dimension members, which can greatly decrease the size of your local cubes and the time needed to create them.

3. When creating local cubes with CubeSlice OLAP ASSL, you have the option of changing the key attribute of each dimension. This also greatly decreases the size of your local cubes and the time needed to create them.

4. Local cubes created with CubeSlice OLAP ASSL can have slower browsing performance when certain features are included in the local cubes. Here are the features that can slow browsing in a CubeSlice OLAP ASSL local cube:

  • Unary operators
  • Custom rollup formulas
  • Measures with Distinct Count aggregation

5. Local cubes created with CubeSlice OLAP ASSL are sometimes (but not always) larger than local cubes created with CubeSlice Relational ASSL.

6. The creation of a local cube from a server cube containing linked dimensions or linked measure groups is supported by the CubeSlice OLAP ASSL. It is not, at the present time, supported when using CubeSlice Relational ASSL.

7. More cubes can be successfully created with slicing when using CubeSlice Relational ASSL. Local cube creation with CubeSlice OLAP ASSL can fail when slicing on the time dimension, if there is also a semi-additive measure included in the local cube. Slicing can also cause local cube creation to fail when a many-to-many dimension is included in the local cube. CubeSlice Relational ASSL does not have these problems.

8. When unary operators are used along with measures using semi-additive aggregations, the data in the local cubes can differ. This happens in the following situation:

  • Unary operators are being used
  • A semi-additive measure (such as Most Recent Non-Null Value) is being used
  • There is no data for a time period in the source data, so that a previous value is displayed in the cube.

In this situation the CubeSlice Relational ASSL local cube ignores the value in the semi-additive measure and does not include it in aggregations, matching the behavior in the Analysis Server cube. The CubeSlice OLAP ASSL local cube uses the value in the semi-additive measure and aggregates it into the higher levels using the specified unary operator.