January 21, 2008 at 8:30 am
i would like to increase the speed at which it takes some of the cubes in analysis services
to process.
if i add indexes to some of the tables in the relational database... shouldn't that improve the
speed of cube processing?
if thats true... how do i know which tables to target for index creation?
thanks in advance.
_________________________
January 22, 2008 at 9:17 am
I did a couple things to improve our cube processing time.
First, during the cube process I ran a profiler on our data source to uncover the queries that were being run to retrieve the data. You can also view these queries when you are manually processing the cube from Visual Studio or SSMS. I then threw those queries into SSMS and reviewed the query plans. I updated indexes accordingly which really improved processing times.
The second thing I did was implement a process add for our measure groups and one of our dimensions. That way instead of processing the entire set of data, only that data that was added was processed. The process add for the dimension is not available through the GUI and must been done with a rather lengthy XMLA script, but it did cut down our cube processing time.
Hope that helps!
January 22, 2008 at 11:46 am
it does. appreciate it. 🙂
_________________________
January 23, 2008 at 2:19 pm
I found if I indexed all my keys to the dimension tables (25 or so), that my cube processing time dropped substantially.
Good luck
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply