Please start any new threads on our new site at https://forums.sqlteam.com. We've got lots of great SQL Server experts to answer whatever question you can come up with.

 All Forums
 SQL Server 2000 Forums
 Analysis Services (2000)
 OLAP Cube Refresh mechanism where fact table is frequently updated

Author  Topic 

AskSQLTeam
Ask SQLTeam Question

0 Posts

Posted - 2002-01-04 : 09:31:16
Siva writes "Hi there,

My application is having one cube where the fact table is frequently updated. Fact table contains 20 million rows. The Fact table is Partioned View. Is there any other process apart from refresh data option of analysis Server to get the effect of updated rows.

AS of now there is no partiotions in the Cube. And Cube process option is refresh data. It is taking 4 hrs to process.

I am searching for a solution where i can reduce the Process to less than 1 hrs.
If anyone has any ideas will be greatly appretiated.

thanks in advance
Siva"

Tim
Starting Member

392 Posts

Posted - 2002-01-29 : 17:04:03
The first thought I have is that your speed problem may be more likely to do with your partitioned view. Theusual approach is to have an ETL process that gets the data from different databases/servers into a single data mart. Then your fact table will be a table and I am sure will be a better performer.

Second thought is to make sure that you have appropriate indexes on your tables.

Finally, you can optimise the queries that AS generates when building cubes (assuming you have a good clean star schema). By default it can do a lot of unecessary joining. Some basics:

- query complexity will be reduced if you use unique keys at the leaf levels of your dimensions
- use optimize schema command to reference member keys from the fact table rather than joining dimension tables

If you have the right design, the query behind your cube processing should get rows from the fact table only. It won't have to join dimension tables to your partitioned view.

HTH

Go to Top of Page
   

- Advertisement -