Kbase P163321: Norwegian characters do not sort correctly with UTF-8 database
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
4/7/2010 |
|
Status: Unverified
SYMPTOM(s):
Norwegian characters do not sort correctly with UTF-8 database
The sorting of the data looks like:
AA
Duck
Smith
Walker
Å
Ås
Æ
Ær
Ø
Ødegaard
This is a basic ASCII sort and is not correct for Norwegian. In addition to the general sort sequence being incorrect, in Norwegian AA = Å (and aa = å) and so should sort the same.
The 4GL client is running a query that is being resolved by an index on the database, but the results are not displayed in the correct order.
FACT(s) (Environment):
The correct Norwegian sorting of the data should be:
Duck
Smith
Walker
Æ
Ær
Ø
Ødegaard
Å
AA
Ås
The database uses the UTF-8 code page and basic collation.
The 4GL client session uses -cpcinternal UTF-8 -cpcoll basic.
In Norwegian AA = Å and so should not be sorted as the ASCII string "AA".
All Supported Operating Systems
OpenEdge 10.x
CAUSE:
The database is using the wrong collation.
FIX:
For Norwegian sorting of data the correct collation would be the ICU collations ICU-nn (Norwegian Nynorsk) or ICU-nb (Norwegian Bokmal).
To correct the problem load the database with the ICU-nn or ICU-nb .df file in the prolang directory. After doing this and rebuilding the indexes the database brokers and client sessions should start with -cpinternal UTF-8 and -cpcoll ICU-nb (or -cpcoll ICU-nn).