If multiple generations
of a cube have been created, special considerations are needed when
archiving or deleting these generations. It is possible to archive
an old cube generation into another schema or under another name.
The cube will then be available for access to historical data. If
you are deleting cube generations, it is important to note that a
cube that is an earlier generation must be deleted with PROC OLAP.
Deleting only the metadata will not work for generations of cubes.
Because cube generations
generally look through to aggregation data in previous generations,
cube data cannot be physically deleted as long as another generation
is dependent upon it. Only those physical files that are not needed
by the new cube (for example, member and property trees) are actually
deleted. The aggregation files remain available to subsequent generations.
In addition, the metadata registration in the metadata repository
is deleted.