Many thanks for clarifying this point.
So when do we use different databases for a cube. If we are building
analytic apps in a company invariable at some stage the same dimension
will be needed in another cube if we split them into different
databases we will risk maintaining the same dimension is 2 places.
Problem of building all cubes in one database is archiving the database
can be very time consuming.
Can you please share the dos and donts on when should we use a
different database is it for every data mart or when we expect
dimensions in a new cube is already in a existing database then do we
build the cube in that database.
If we try building a new cube where there are some dimensions in the
existing database but if the existing AS database as a source system
from a SQL database and the new cube is from a different SQL database
or data source how do we manage these kind of scenarios.
Thanks
Karen
Deepak Puri wrote:
Quote:
The safest approach would be to update the dimension after each fact
table is loaded.
Not sure why you would need to share a dimension across 2 AS databases,
rather than using a single database, but each database could have its
own copy of the dimension, while using the same source table/view. Of
course, this would mean that the dimension would have to be processed
separately in each database.
- Deepak
Deepak Puri
Microsoft MVP - SQL Server
*** Sent via Developersdex http://www.developersdex.com *** |