Character data is placed in the database using the exact binary representation that is passed from the application.
This usually means that character data is stored in the database with the binary representation of the character set used by your system. You can find documentation about character sets in the documentation for your operating system.
On Windows, code pages are the same for the first 128 characters. If you use special characters from the top half of the code page (accented international language characters), you must be careful with your databases. In particular, if you copy the database to a different machine using a different code page, those special characters are retrieved from the database using the original code page representation. With the new code page, they appear on the window to be the wrong characters.
This problem also appears if you have two clients using the same multiuser server, but running with different code pages. Data inserted or updated by one client might appear incorrect to another.
This problem also shows up if a database is used across platforms. PowerBuilder and many other Windows applications insert data into the database in the standard ANSI character set. If non-Windows applications attempt to use this data, they do not properly display or update the extended characters.
This problem is quite complex. If any of your applications use the extended characters in the upper half of the code page, make sure that all clients and all machines using the database use the same or a compatible code page.