NDO Schema trees can be created in any system encoding. For NCF Schema loading, it is advisable to create NDO schema trees in the native system encoding. The database drivers and the database server software perform a character set negotiation when the schemas are loaded to the database. The database drivers use the native system encoding to establish their end of the negotiation. Attempting to load data of another character encoding set can result in improper translation. For NCF schema loading, the e-ADK examines the NDO schema tree and, if not in the native system encoding, converts it. Characters that are not convertible do not result in an exception being thrown. Instead, the non-convertible character is replaced with a fill character which will result in improperly loaded schema information.
NCF format prefixes must be ASCII-compatible single-byte characters. NCF format names, however, can be crated using multi-byte characters. The formatter database enforces a 120-byte maximum length for format names. If your format name is longer than 120 bytes and contains multi-byte characters, you must set the configuration key I18N_Condense equal to true. The condense routine truncates characters from the string. Setting I18N_Condense equal to false uses the old condensing algorithm which only works with single-byte data.
