Unicode support for federated systems

The database in the Unicode code page provides federated server environments that are platform independent. The database can manipulate data that is stored in various code pages on different data sources.

In Figure 1 a company has branch offices in different countries. Each branch office stores customer data with its own databases in their own code page. The Microsoft SQL Server database stores data in code page A. The Oracle database stores data in code page B. Code page A and code page B are in different territories. To integrate the data from the different territories, the company can set the federated database's code page to Unicode. The company can then join the tables to see the total number of purchase orders, regardless of territory.

In some cases, when the data source database code page is SBCS or DBCS, and the federated database is Unicode, the text string will be truncated during conversion. This is because the federated server performs the conversion based on character length rather than byte size. For example, when the federation server is UTF-8, the data source database works in IBM970. For the character "가 ", the character length is 1, and the GBK byte is b0a1, the byte size is 2. When converting to UTF-8, the byte is eab080, the byte size is 3. The federated database only gets the first byte, not all three bytes. Increasing the length of the nickname column can get the entire result.