I think I must have confused you, or I wasn't clear in my previous statement. So, with that being said, I'll try to re-state:
A non-clustered index is less efficient than a clustered index as SQL Server must traverse the various index pages to locate the record's location. Then, SQL Server must go and find that record, either by using a clustered index if one exists, or the RID.
If you don't have a clustered index on a table, but you do have a non-clustered index, I don't think you'll see much of a performance difference between using an RID versus a clustered index in one existed.
I hope that makes my previous statement a little clearer...
>>Larry,
>>
>>I didn't mean to suggest that any performance difference would crop up using a non-clustered index without a clustered index. I was simply pointing out that a non-clustered index would be far less efficient than a clustered index.
>>
>
>Okay Travis, I'll bite. How do you determine better efficiency if not via better performance? It's not like we can compare the wear and tear on a piston under two different conditions.
>
>I realize the performance boost wouldn't be apparent in all cases. For instance, when the number of rows is relatively low, it may not be. However, at some point, if the process is more efficient, it must translate into better performance.
Travis Vandersypen