February 27, 2006 at 12:04 am
I just want to ask that once a Cube is processed and deployed, is there any need for the data stored in the "AdventureWorksDW", as the cube stores its own copy of the source data, so it means that we can delete the data from its source database.
Secondly how and when the cube collects the new data in the source, whether it collects the changed or new data or it collects or the data and replaces its previously collected data. If it collects the changed or new data....then will it be an scheduled task or we manually have to process the cube again.
February 27, 2006 at 3:31 am
In the normal MOLAP storage scenario it is true that the data from the facts tables (or at least only the relevant columns that are used for measures and keys that link to dimensions used in the cube) gets stored in the cubes storage. It is possible to dispose of the data in the relational source and still access the details using DRILLTHROUGH in MDX. However, if you are planning to adjust the cube's structure in the future, i.e. add additional measures or dimensions (which happens in many cases), than you will inevitably have to Reprocess the cube, its measure groups or partitions. This will clear all the data retained and will try to source it again from the relational data source. In other words I would not recommend getting rid of the data from your DW unless you have very good reason for it (perhaps limited storage space) and you will keep the cube structure unchanged (this doesn't apply to changing calculated measures or scripts and doing some minor changes to dimensions). In my opinion it is better to view OLAP cubes as a convinient way to cache and access your relational Data Warehouse data. The cache is not a very good place to store the only copy of your valuable data.
As for the second question, you have a multitude of choices: you can push new data in you cube from SSIS package, perform incremental updates of partitions or reprocess partitions. Have a look at this article: http://msdn.microsoft.com/sql/default.aspx?pull=/library/en-us/dnsql90/html/sql2k5_asprocarch.asp It describes processing architecture of SSAS 2005 and choices you can have depending on your requirements.
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply