Database Character Set and Atlas Hub

Atlas Hub stores UTF-8 data in Oracle schema users. Atlas Hub does not use the Oracle conversion mechanism. It performs the conversion to UTF-8 itself.

Because Atlas Hub does the conversion, it is very important to prevent Oracle Client from also converting the data. A double conversion will result in garbage.

Oracle software does the conversion only when the Oracle Client’s character set (specified by NLS_LANG) does not match the database instance’s character set. Therefore, it is very important that the client and server’s settings match. Atlas Hub configurations require that Oracle database instances use the WE8MSWIN1252 character set. If your Atlas Hub database was previously Oracle 8i, then Oracle database instances use the WE8ISO8859P1 character set. Atlas Hub also supports AL32UTF8 character set.

The Server Administration tool does not allow you to create a new Atlas Hub server configuration unless the underlying database instance’s character set is the correct character set. When Atlas Hub starts, it also ensures that the database character set matches the characters set portion of NLS_LANG.

Due to the numeric setting differences between different languages and territories, Atlas Hub also does not start unless AMERICAN_AMERICA (the US numeric setting) is the setting for the language_territory portion of NLS_LANG. Failure to do so causes exceptions for Atlas Hub. This does not mean that numbers and so on will be displayed incorrectly. For example, in Germany and other countries, a decimal point is represented as a comma. Such numbers will still display correctly.