IBM Cognos, Руководство пользователя

Страница: 193 / 411

This action removes all categories not created or changed since your specified
date.
3.
Regenerate the cube and confirm that the stale-dated categories are no longer
present.
Modifying a PowerCube
If the business information requirements or data access rights of your report users
change, you will likely need to make corresponding changes to your PowerCubes
or cube groups.
Update an Existing PowerCube
If your changes are minor, you can use the Update PowerCube feature ( Run
menu), instead of deleting and replacing the entire cube, to update
v
object names, short names, labels, and descriptions
v
security objects and their associated views
v
currency conversion tables
Omit Dimensions Used Only for Detail Reports
If your model has partitioning on detail dimensions, querying at the lowest (leaf)
level will be slow because reporting must be done across multiple partitions. To
avoid this problem, you can create a separate PowerCube that omits dimensions
used only for lowest-level detail reports. The resulting cube is smaller, quicker to
build, uses smaller partitions, and processes summary queries more quickly.
You can choose to omit dimensions from a cube, even if it uses auto-partitioning.
For more information, see “Omit Dimensions from a Cube” on page 132.
Update Cubes Incrementally
It is best to place frequently updated data in one or more separate data sources, so
that you can perform incremental updates to quickly add the latest changes to
your existing cube.
Tip:
Make sure that your data sources contain only new data; otherwise, data
duplicated in your source files will be duplicated in the cube.
Incremental updates can save processing time. However, you must consider the
following:
v
The first time you run an incremental update, include all the categories for the
levels used for partitioning. This ensures the best performance from an
incrementally updated cube.
v
If you add new categories to a partition level in a model, the new categories are
stored in the parent partition level.
v If you add a large number of new categories, the access time for report users
may be adversely affected because of the increased size of the summary cube.
Also, over time, if new categories are introduced, the incremental updates may
reduce the effectiveness of the partitioning scheme in the cube, which degrades
performance. In either of these cases, we recommend that you repartition your
model.
v
After the first incremental update, Cognos Transformer reverts to using the
less-effective
Categories method, rather than the Auto-partitioning method.
Chapter 8. PowerCubes in Production Environments
179