Re: [Firebird-odbc-devel] Similar Characters in querys

[Peter J. <pj...@wa...>]

Jorge Andr=E9s Brugger <li...@da...> wrote:
> Well, thanks for your interest, but I think is a little more 
> complicated.[...] The problem is that I dont know if the accented charac=
ter is the last
> of the string used for querying, or not. Do you understand me?

To put it bluntly, you didn't try my suggestion and you don't know
how multi-level collations work.

Using this test script:
set names ISO8859_1;
create database 'C:\es_es2.fdb' user 'sysdba' password 'masterkey';
create table t1 (
    c1 char(8) character set ISO8859_1,
    c2 char(8) character set ISO8859_1 collate ES_ES
);
insert into t1 (c1, c2) values ('Paaa', 'Paaa');
insert into t1 (c1, c2) values ('Puuu', 'Puuu');
insert into t1 (c1, c2) values ('PEPO', 'PEPO');
insert into t1 (c1, c2) values ('pepo', 'pepo');
insert into t1 (c1, c2) values ('Pepo', 'Pepo');
insert into t1 (c1, c2) values ('Perejil', 'Perejil');
insert into t1 (c1, c2) values ('P=E9rez', 'P=E9rez');

select * from t1 where c1 BETWEEN 'pe' and 'pezzz';
select * from t1 where c2 BETWEEN 'pe' and 'pezzz';


The result from the first select, which uses a column
with codepoint collation is:

C1       C2
=3D=3D=3D=3D=3D=3D=3D=3D =3D=3D=3D=3D=3D=3D=3D=3D
pepo     pepo


But the result of the second query, which uses a
column with multi-level collation, is:

C1       C2
=3D=3D=3D=3D=3D=3D=3D=3D =3D=3D=3D=3D=3D=3D=3D=3D
PEPO     PEPO
pepo     pepo
Pepo     Pepo
Perejil  Perejil
P=E9rez    P=E9rez

As you can see, the item with the accented e is
selected, which is easier to understand when
you do some tests how ES_ES sorts
strings.

Best Regards,
Peter Jacobi