Michael, the character set is set to UTF8. Frankly I think this is most likely a bug/weakness in the UDF.

--- In firebird-support@yahoogroups.com, Michael Ludwig <milu71@...> wrote:
>
> dancooperstock schrieb am 01.06.2012 um 16:27 (-0000):
> > I can force the error with just this SQL:
> >
> > SELECT f_GenerateSndxIndex('ÉLISE') FROM rdb$database
>
> Would using an introducer help? Along the lines of:
>
> SQL> show table z;
> A VARCHAR(20) CHARACTER SET UTF8 Nullable
> SQL> insert into z values ('Übel');
> Statement failed, SQLSTATE = 22000
> Malformed string
> SQL> insert into z values (_win1252'Übel');
> SQL> commit;
> SQL> -- all characters in good shape
>
> That's an ad-hoc conversion of literals.
>
> Also, if in isql, consider using an option such as -ch win1252 to
> specify a character set your UDF can live with. Same for programmatic
> access. For truly multi-alphabet data like Greek and Russian together
> in one column you need Unicode, of course.
>
> Michael
>