Hello Martijn,

> Just had a try with ISO8859_1, collation DU_NL -
> when I ORDER a column with these values:
>
> And
> AND
> aNd
>
> it ens up like:
>
> aNd
> And
> AND
>
>
> While using collation ISO8859_1 (the default), I get:
>
> AND
> And
> aNd

Well I never !!

I often use collation DU_NL in domain and field definitions, but
because I "knew" (ahem) it was case-sensitive, I also take additional
measures so that users can sort a column case-insensitively where
needed. Although easy to implement in IBO (just specify NOCASE) there
is of course a performance penalty.

Now I just sorted a DU_NL column with names in isql and yes:

...
Ufuk
Ugbad
Ugur
uiuiui
Umit
Valter
...
Zulay
Zulfu
zwap

(In case anyone wonders who the heck would call his child *uiuiui* or
*zwap*, I just added those for the check :-))

But also if I specify "order by voornaam collate en_uk" or even
"en_us" the sort is case insensitive. Collation iso8859_1 is case
sensitive like you said.

I tried this under Linux and Windows: same output.

I'm using Firebird 1.0, don't know if the same goes for IB.

BTW: Accented words fit in nicely too, ie O-umlaut counts als O in the
DU_NL collation, but that was something I already knew (this was my
main reason for choosing that collation).


Greetings,
Paul Vinkenoog