>the primary key index does not have to be clustered. I am glad you brought this up,though. Since I use Strataframe's Data Deployment Toolkit ( which is an SMO based schema wrangler ) I handle my keys there and didn't realize that in SSMS if you create a primary key it is in fact automatically clustered.
>
>To change this :
>

My rule of thumb is to only do this in INSERT-heavy tables. [The numbers I've read are possibly pure junk today
as more RAM and faster disks change performance patterns, but:] If pure inserts make up more than 25% of
all accesses I think about it. Otherwise the technical grounds Mike cited should give an edge to clustered...

my 0.02 EUR

thomas