cyrillic sort order, ru_ru problem
nekoj needed a database with Macedonian Cyrillic sort order on IDS 9.30 UC2 on Linux; ru_ru worked but sorted incorrectly, and the only mk_MK entries were in the cm3 registry file. scottishpoet suggested writing and compiling a custom locale. david advised reading the GLS manuals, noting IDS 10 has Unicode support, and raising an IBM support call. Unresolved.
Auto-generated by Claude from the posts below — may be imperfect; read the full thread.
Hi group, I need to create a database with cyrillic (macedonian) support. I have manager to do it with ru_ru support and I can read/write to it no problem. The point is that our alphabet has diferent sort order than russian so some sorts look the way they should not. I have been digging a while trough the documentation and this forum but I am still confused. The only mk_mk string under gls directory is $INFORMIXDIR/gls/cm3/registry #mk_MK.ISO8859-5 915 # 0x0393 IBM AIX locale (os/ibm/mk_MK.ISO8859-5) #mk_MK 915 # 0x0393 IBM_AIX locale (os/ibm/mk_MK) what is the meening of this. On the other hand i have been searching trough ru_ru directory and found out few .lco files. I assume that LC_COLLATE can/should be set here so that every letter in the alphabet get its own "weight factor"? Am I right or wrong about this. Any idea how to find a workaround and solve this problem? 10x a lot aleksandar
Not sure but you may have to write and compile your own locale nekoj wrote: > Hi group, > > I need to create a database with cyrillic (macedonian) support. I have > manager to do it with ru_ru support and I can read/write to it no > problem. > The point is that our alphabet has diferent sort order than russian so > some sorts look the way they should not. > I have been digging a while trough the documentation and this forum but > I am still confused. > The only mk_mk string under gls directory is > $INFORMIXDIR/gls/cm3/registry > > #mk_MK.ISO8859-5 915 # 0x0393 IBM AIX locale > (os/ibm/mk_MK.ISO8859-5) > #mk_MK 915 # 0x0393 IBM_AIX locale (os/ibm/mk_MK) > > what is the meening of this. > On the other hand i have been searching trough ru_ru directory and > found out few .lco files. I assume that LC_COLLATE can/should be set > here so that every letter in the alphabet get its own "weight factor"? > Am I right or wrong about this. > Any idea how to find a workaround and solve this problem? > > 10x a lot > aleksandar
which IBM version? IDS 10 has the Unicode support from DB2! Very few people seem to use GLS stuff First read the GLS manuals then I would raise a support call with IBM.
How to compile my own locales? is glfiles somehow related? Version: IDS 9.30 UC2 on Linux. meanwhile I may try the 10 version but it may be only for wducational and developement porposes, no upgrade of the production servers in plan so far. thank you all, Alex