In general, the content columns of external files system tables are treated as image. However, special conversions are performed when a content column is assigned to a Unicode column, i.e. a column of datatype unichar,univarchar, unitext, or java.lang.String. Such assignment of a content column to a Unicode column occurs in the following contexts:
An insert command used to insert a Unicode column from a subquery that references a content column.
An update command used to update a Unicode column with a new value that references a content column.
A convert function call that specifies both a target Unicode datatype and a source value that is a content column.
In assigning a content column to Unicode, use these rules:
If the source document has a BOM (Byte Order Mark), to convert the source document the BOM must indicate UTF-8 ir UTF-16. If the BOM indicates UCS-4, an error is raised. UCS-4 is not supported.
If the source document has an XML header that includes an encoding clause, but no BOM, the encoding clause must specify the server character set or UTF-8 to convert the source data. An encoding clause that specifies a character set other than the server set or UTF-8 raises an error.
If the source document has no XML header, a header with no encoding clause, and no BOM, the processor treats the character set as UTF-8, and converts the source data.
If an error occurs during a conversion, an error is raised, but the statement continues.
