Depending on the business
needs for a cube, it might be more suitable to update a cube in-place.
A separate version of the cube is not created. The process of updating
in-place enables you to add data to a cube while that cube is still
online. The cube maintains its name and OLAP schema assignment. This
process might be more appropriate for some users and their reporting
needs, as it is less complicated than creating incremental generations
of a cube. However, there is no opportunity to test and verify the
update in a separate process.
You can update a cube
in-place by using the UPDATE_INPLACE PROC OLAP option. This option
is used with the ADD_DATA option to signal that an update will occur.
Here is an example of the UPDATE_INPLACE option:
proc olap data=mylib.newdata cube=cubeA add_data
update_inplace;
metasvr host="myhost" port=8561 protocol=BRIDGE repository=my repository
olap_schema=prodSchema;
run;
The process of updating
in-place does not create a new metadata registration when the update
is complete. The existing metadata registration points to the new
cube generation and the old generation is deleted. In addition, you
do not have to disable the cube prior to beginning the in-place update.
OLAP server sessions referencing the old cube will continue to do
so until the last session closes the old cube. After that, new sessions
will begin to reference the updated cube. However, it might be desirable
to have a specific “cross-over” to the new cube at some
point after the update completes.
See
Update a Cube In-Place for a discussion of
updating a cube in-place with the SAS OLAP Cube Studio Incremental
Update Wizard.