[unixODBC-dev] Question about who does the localisation conversions

Nick Gorham nick at lurcher.org
Tue Mar 1 20:11:12 GMT 2005

Bill Medland wrote:

>Can anyone help on this?
>We have a program running under Wine through unixODBC down into IBM DB2.
>In order to get Spanish display under Wine we set the LANG to es_ES under 
>Linux.  (Under Windows we set the regional settings)
>When we get a decimal number (DECIMAL length 19 scale 3) from the database 
>under windows through ODBC it comes through with a period as the decimal 
>point separator.
>If I use the db2 command itself to query the database, using 
>LANG=es_ES db2
>then it displays with a period for the decimal point.
>If I use isql to query the database, using
>LANG=es_ES <dbname> <username> <password>
>then it displays with a comma for the decimal point.
>Can anyone confirm or deny for me that unixODBC will be doing the conversion 
>from the underlying database type on the basis of the LANG setting?  Or is it 
>likely that it is the actual database drived?

unixODBC will try and use iconv if its present to convert to and from 16 
bit unicode to ascii if needed, but will not do any of the sort of 
conversions you are talking about.

I would check if there are any attributes that can be set either via the 
API or in the ini file for the db2 driver.


More information about the unixODBC-dev mailing list