Collation for a new MS SQL Server
Posted on 2013-01-09
We're about to install a new MS SQL Server which would hold all our new DW and OLAP databases for the company.
We intend to insert into it data from legacy DBs, which, among other things, hold URLs with unicode characters.
In general we would like to be able to support many kinds of non-English languages, but we're not sure which collation to choose for the server setup and for the instances.
Our main candidates are:
1) utf8_general_ci - which would have been the best option maybe, but it seems that you cannot assign this collation at server-level! Problem is, if we would define it at instance level, what do we assign to the server-level, and what is the affect of having them not the same?
2) Latin1_General_CI_AS - the fallback of the above maybe, since it can be defined at both the server and instance levels, but what do we do with our unicode requirements?
3) Finnish_Swedish_CI_AS - since we have some swedish and finnish sites, some people refer to this collation is maybe better against the "latin" one above.
So the end question is: what collation should we choose at server-level and on instance-level?
Thank you in advance!!!