September 17, 2014 at 1:19 pm
0I need some advice about the design if the OLAP dimension tables in order to implement hierarchies in Analysis Services.
I'm trying to keep as much as possible to a Star schema, but I also want to implement hierarchies in Analysis Service to ensure query performance is optimal.
My issue is that the hierarchy I need spans multiple tables which all link to my fact separately like a star schema so if we want to query the relational database directly the performance is good.
I wanted to know if adding a linking key column between the tables so they can be also joinned together like a Snowflake and then telling Analysis Services about both joins and then building the Analysis services heirarchies is the best approach. Not sure how Analysis Service will use the joins i.e. which ones will it use. Would it use one join for the hierarchy and another for the other attributes.
or
Should I just join them together like a Snowflake and only tell Analysis Service this is how to join the tables.
or
Should I flatten the tables into one and use that. Hopefully this is not the right approach as there are over 8 tables with many rows over, one dimension has 30,000 rows.
Viewing 0 posts
You must be logged in to reply to this topic. Login to reply