Data in the main language are stored in the original object type "translated" table using the database character set. Data in the supplementary languages are stored in the additional "translation" tables using unicode column type.

Example:

Oracle database character set is NLS_CHARACTERSET = WE8MSWIN1252. The '€' symbol is thus stored using 1 byte in a column of VARCHAR2 datatype.
Oracle national character set is NLS_NCHAR_CHARACTERSET = AL16UTF16. The '€' symbol is thus stored using 2 bytes in a column of NVARCHAR2 datatype.
The maximum column size for NVARCHAR2 is 4000 bytes which corresponds to 2000 characters in AL16UTF16. Therefore the maximum length of a main language column in the database table is twice as big as the maximum length of data in the translation table.

Main language of data in the standard and customer databases

Standard USU databases have English (en_US) as the main language of data. This is because English characters are represented in all Oracle database character sets and can be transported to any platform.

When the main language of data in the customer (target) database differs, the difference is handled in the following way:

• If the target language of data is the same as in the source (standard) database:

  Database import creates records as they are in the source database

• If the target language of data is different from the main language in the source database:

  Database import saves values in the target language in the translated table and values in the source language in the translation table.

  If no value in the target language exist, then the value in the source language is also stored in the translated table as a fallback value with corresponding encoding mapping.

• If the target language of data is not defined yet:

  Database import creates records as they are in the source database.