Hi all,
Our databases which are deployed at client sites have somehow ended with a mixture of column collations. The 'fix' the developers have been putting in is to use COLLATE database_default when comparing columns.
I suspect this affects performance - particularly index selection?
I am looking to write a script to set all columns to the database collation - to do this I will:
1. Find all columns where the collation is different to the database collation
2. Find all indexes on columns found in step 1
3. Drop the indexes found in step 2
4. Change collation of the column
5. Recreate the indexes
I tried disabling the index, as I expected, this does not allow the collation to be changed.
Is there some option other than scripting the indexes out?
Also - appreciate any comments on how performance is affected with different collations...does it matter how close the collations are e.g. Accent Sensitive vs. Accent Non Sensitive when no accents are used? I doubt SQL Server considers this?
Thanks for reading,
Clay