The character set of the schema the function is defined in. That keeps it least confusing and easier for the users to understand.
On 4/3/07, Daniel John Debrunner <firstname.lastname@example.org> wrote:
> 7)For user defined functions' that return character string type, the return type's collation
> will have the same collation as current schema's character set.
The "current schema's" character set, or the character set of the schema
the function is defined in? And if it's the current schema, is it
current at function definition time, or current when used?