In the SQL database for GP, it seems there are several large tables that have an identity column, yet no clustered index. Is this by design, or something that is just left up to the admin to create?
*This post is locked for comments
In the SQL database for GP, it seems there are several large tables that have an identity column, yet no clustered index. Is this by design, or something that is just left up to the admin to create?
*This post is locked for comments
Hi,
Clustered indexes are used for specific tables in GP. you can refer the GP architecture document link related to clustered indexes in page 7
There is no need to add custom indexes most of the time as GP already has these created. Could you give an example of a table you are wondering about?
Well, this helped answer a bit.
www.simple-talk.com/.../should-you-always-use-a-clustered-index
The key here is that a table with a clustered indexes AND non-clustered indexes, a given query will work differently than having only non-clustered indexes.
With non-clustered and no clustered - the non-clustered indexes have pointers directly to the rows on disk, where as if there is a clustered index, the pointers in the non-clustered index point to locations in the clustered index...which then point to the rows on disk.
Clear as mud ? :)
André Arnaud de Cal...
292,494
Super User 2025 Season 1
Martin Dráb
231,305
Most Valuable Professional
nmaenpaa
101,156