Unanswered: problems with collate and scandinavian languages
I spend quite some time trying to make collate to work properly on a UTF8 DB with scandinavian languages (more precisely danish) but I can't make it work properly.
I am using DB2 8.1.6 on RedHat 9 linux (x86). I create a DB like this:
db2 create db test collate using system
then I connect to the db and check with GET DB CFG that the DB is UTF8 (code page 1208), territory DA. Everything is just fine. I create a simple table with few rows and then try to sort them but the result I get is not correct. The problem is with the special danish letters 'æ','ø' and 'å' are not properly sorted. The result I get is like if IDENTITY collate function was used which cannot be used for languages like danish where strings cannot be compared by comparing code points because that does not give a proper alphabetical order. Note that the code points for the special characters have code points beyound 0xFF. When I created the DB I have specified SYSTEM collate to be used. I would expect that DB2 will build collate table with proper weights obtaining the information from libc and the current locale when the DB is created. But things do not work as expected.
Am I doing something wrong? Do I need setup a custom collate weight table (which would involve using the CLI interface) to make this work properly or there is another standard way?