Hi everyone,
I am running a complex reporting system on SQL Reporting Services and I think that analysis services may provide some benifits, however as I am new to it, i could use a little guidance.
Basically I am reporting on a huge volume of very granular information. Because there is so much data we prune the system such that at any point it only reflects 30 days worth of information. Because of this, I can't really run reports to look at statistical information or trends over time (say 6 months to a year).
I was wondering, would it be possible to take certain information from the database before it is pruned and store this seperately to be used by Analysis services to report on long term trend data? Can analysis services do something like this?
Any help or advice would be appreciated!
Kind regards
Taz
Analysis services will use your detailed information only when the cube is processed. After that the cube is a separate, multidimensional structure with aggregations designed above your data source leaf level.
Be aware of a problem with designing cubes from report systems. Do not use more than 10-15 dimensions in the cube. Report applications can often have more than this.
I recommend building a star schema. You can see that in the Adventure works DW db that is part of the samples in the SQL Server 2005 installation. You can also find information about star schemas here(www.kimballgroup.com). Star scemas can be built by using views.
HTH
Thomas Ivarsson
No comments:
Post a Comment