I do not know what your case exactly is but next two proposes are exposed:
One solution is to have different databases with different collations (one for each language you need).
If you need to maintain all your data in one database next will be the right solution:
If you have to store chinese characters, spanish, german, etc. in just one database, one solution (I think the only one) is to convert your character datatypes to unicode:
- Instead of char datatype you have to use nchar datatype.
- Instead of varchar you have to use nvarchar datatype.
- Instead of text you have to use ntext datatype.
nchar, nvarchar and ntext use unicode characters that are saved with two bytes per character instead of one byte. Unicode columns are almost independent of the collation of your database.
On the other hand, query sentences are affected. For exemple, you will have to use:
insert into table1 values (N'hello') instead of: insert into table1 values ('hello')
However, globalization must be studied carefully. I recommend to investigate more before you decide the final solution.
Here you have a good link to start to know more about unicode: