SQL Kiwi (1/31/2012)
peter-757102 (1/31/2012)
I took a look at the plans and what confused me is the apparent lack of clustered indexes (i noticed table scans instead of clustered index scans)!The table names made me think these are fact and dimension tables in a data warehouse. It is not uncommon to use heaps in that scenario.
Can you explain (or have a link that does a good job at it), why one would want to go that route over clustered storage in that scenario?
The one thing that comes to my mind is reduces storage for simple entities, but when you start to add indexes and query it like it is relational, that small 'advantage' evaporates real fast.