Richard's script seems to confirm that something is buggy, so I hold fire
trying to produce something else that only shows the same thing.
Staffan
On Mon, May 6, 2013 at 5:00 PM, Richard Hipp <d...@sqlite.org> wrote:
> On Mon, May 6, 2013 at 10:56 AM, Simon Slavin <slav...@bigfraud.org>
> wrote:
>
> >
> > Unless you include your collation code, it's hard to reproduce your
> > behaviour. Can you please try to find a demonstration with a collation
> > order that's native to SQLite ? Perhaps a three-row example using
> "COLLATE
> > NOCASE" ?
> >
> >
> You are correct, Simon, that Staffan's message was really just a vague hint
> of something wrong. But based on that meager hint, I have been able to
> come up with a suspicious-looking test script using the built-in "nocase"
> collation. To wit:
>
> CREATE TABLE t1(a);
> INSERT INTO t1 VALUES('abc');
> INSERT INTO t1 VALUES('def');
> CREATE TABLE t2(a);
> INSERT INTO t2 VALUES('DEF');
>
> SELECT a FROM t1 EXCEPT SELECT a FROM t2 ORDER BY a;
> SELECT '-----------------';
> SELECT a FROM t1 EXCEPT SELECT a FROM t2 ORDER BY a COLLATE nocase;
> SELECT '-----------------';
> SELECT a FROM (SELECT a FROM t1 EXCEPT SELECT a FROM t2)
> ORDER BY a COLLATE nocase;
>
> --
> D. Richard Hipp
> d...@sqlite.org
> _______________________________________________
> sqlite-users mailing list
> sqlite-users@sqlite.org
> http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
>
_______________________________________________
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
