May 7, 2007 at 8:47 am
I have a cube with about 20 Measure Groups and 30 dimensions. I built some dimensions using BI Development Studio a few weeks ago. The aggregations were build for any measure group without problems.
Now I changed some attributes and want to run the aggregation wizard on my test plattform to design new aggregations. But there were no aggregations build. It shows 0 aggregations.
I read that sometimes this happens when the cube is too complex. But the complexity did not grow very much.
What should I do to be able to design aggregations again?
Best regards,
Stefan
SK
May 7, 2007 at 11:40 pm
Is this AS2000 or AS2K5? If the latter, then you'll need to ensure that your attribute relationships are set up correctly (ie it's not enough to simply have user hierarchies - google Elizabeth Vitt for a good article explaining how you can influence the agg candidates). Also check out the performance guide (or again, google on msdn) for info about the 20 (or is it 30) % rule, basically if the storage requirement for the aggs is larger than a given percentage of the size of the cube, it won't consider the agg. NB this is also influenced by the row counts for your partitions, make sure you keep these reasonably accurate.
Steve.
May 8, 2007 at 2:48 pm
Hi Steve,
the attribute relationships are set up correctly, the aggregations design worked the last time and they are build every day. I read the sources. I tried some other measure groups, they worked (aggs designed), it seems that only one special measure group is concerned.
The 1/3 rule does not fit. It should be okay.
The row count for the measure group and the dimension attributes are ok.
Any other ideas?
Best regards,
Stefan
SK
Viewing 3 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply