At 14:32 10.07.2003 +0200, you wrote:

>Thanks, I'm about to get the idea that I have to go to brookstonesystems
>and try to learn how to define my own collation (even though it mentioned
>the frightening letter C, which I've so far managed to steer away from
>using for the last eleven years). Hopefully, the source for the NO_NO
>collation will be available as a starting point.

Hmm, I checked a bit and found that defining my own collation will be
stretching my skills a bit too far if I want to have a summer holiday
within the next couple of years. Though I did find one document (albeit
three years old) saying: "Each InterBase collation is specific to a
character set. All characters in the character set must be sorted. If a
character is not used in a particular culture it can be collated to any
location, but it must still be collated, as it is valid input." This leads
me to believe that IB_SQL is correct in accepting a Swedish character with
Norwegian collation (both NO_NO and SV_SV are collations for ISO8859_1),
whereas my TIB_DSQL errs in rejecting it.

I don't understand why this happens, but have modified the sample
application I was about to produce for Jason when I discovered this problem
and will send it to him (yep, it was another collation problem that made me
aware of this).

Though one thing I did discover, was that RDB$FIELDS.RDB$COLLATION_ID was
not set for a field, even when that collation was specified in the CREATE
TABLE statement. Does anyone know anything about this?

Set