101 reputation
1
bio website
location Hendersonville, NC
age
visits member for 4 years, 8 months
seen Aug 26 '10 at 12:45

Apr
9
comment Changing an MSSQL clustered index field from containing “random” GUIDs to sequential GUIDs - how will it affect indexes
2) The other option would be to generate sequential GUIDs, aka COMB GUIDs. The benefit of this would be that all existing indexes could be kept, and the tables would be physically ordered by the PK. However, as the question states, how will this affect installations that already have lots of data with random GUIDs?
Apr
9
comment Changing an MSSQL clustered index field from containing “random” GUIDs to sequential GUIDs - how will it affect indexes
Yes, I know it is bad, and that is why we want to change it. It seems to me that there are two possible options: 1) Keep the primary key as a "random" GUID, and create a different clustered index, for instance an IDENTITY field in SQL Server. I am a little concerned if this might hurt select performance though, since the main lookup key (the GUID PK) will no longer be used as the physical order on disk, but perhaps this is not really an issue? (continued in next comment)
Apr
7
awarded  Supporter
Apr
7
asked Changing an MSSQL clustered index field from containing “random” GUIDs to sequential GUIDs - how will it affect indexes