Flattening a Dataset to Produce a Rowset

OLE DB Programmer's Reference

It is often useful for consumers to get a flattened view of a dataset. The process of flattening can be illustrated through the following dataset. The axis dimensions are Products, Geography, and Quarters. Slicer dimensions are Measures, Years, and SalesRep.

Flattening this dataset yields the following rowset:

Note   In this table, the names of columns 2–7 are created by concatenating names from the Geography and Products dimensions. It is assumed that the provider fully qualifies the member names from the Geography dimensions to ensure uniqueness—hence "[The World].[Asia]" instead of "[Asia]". However, it is also assumed that member names from the Products dimension do not need to be fully qualified—hence "[Computers]" rather than "[[All] Products].[Product Line].[Computers]". The provider is required to return only unique names. The extent to which the names are qualified to make them unique is up to the provider. A provider need not use qualification as a way of ensuring uniqueness; it can use any method.

1998-2001 Microsoft Corporation. All rights reserved.