Don't know if this question is in wrong forum but I didnt find a better place.
I have a table called "CompanyApplications". The first column is a guid and is the PK (not clustered). The other colum is also a guid and is the company identifier. This is a guid because it comes from another table (the "company" table),
an old table used for other stuff. Then I have a third and forth column for "ApplicationURL" and "ApplName.
My application will do alot of (since 30 000+ users) "select ApplName, ApplicationURL from CompanyApplications where companyId='<the company guid of the user>' order by ApplName). It will make very few inserts into this table. Only when one is
inserting a new application.
I have some questions about this since I know it's not good to have a guid as clustered since it will have SQL Server to always resort the clusted index (because guids are "random"):
1. Is it wrong to have the companyId column as clusted in this example? If so, what is better?
2. I know inserts will be relatively slow when using guid as clustered. But in my case the insert does not need to be fast. Will the slow insert also affect selects from other users. I mean, will select's be slow for other users when the index is resorted
becuase one user make an insert?
View Complete Post