Обсуждение: BUG #1695: "select distinct" makes different output

The following bug has been logged online:

Bug reference:      1695
Logged by:          philip
Email address:      philipguo@yahoo.com
PostgreSQL version: 8.0.3
Operating system:   MS Windows XP sp1 and Linux Fedora core 3
Description:        "select distinct" makes different output
Details:

Hi there,

I installed PostgreSQL 8.0.3 database system on both platforms MS Windows XP
(sp1) and Linux Fedora 3, and both databse settings are exactly the same
with the encoding set to 'UNICODE'. Both database systems are running well
except one thing so far has the different results from those two databases.
Here is a query as the example:

select distinct area from ht_booking.hotels;

The result from the database on Linux FC3 has no problem but the one on
Windows XP. The data output from the later have the duplicated rows. It
seems to make no difference with or without "distinct" key word. BUT, this
happens only under one condition: the data in that column contains chinese
characters.

Please feel free to question me if there is anything I  can do to help with.
Thank you.

> The following bug has been logged online:
>=20
> Bug reference:      1695
> Logged by:          philip
> Email address:      philipguo@yahoo.com
> PostgreSQL version: 8.0.3
> Operating system:   MS Windows XP sp1 and Linux Fedora core 3
> Description:        "select distinct" makes different output
> Details:=20
>=20
> Hi there,
>=20
> I installed PostgreSQL 8.0.3 database system on both=20
> platforms MS Windows XP
> (sp1) and Linux Fedora 3, and both databse settings are=20
> exactly the same with the encoding set to 'UNICODE'. Both=20
> database systems are running well except one thing so far has=20
> the different results from those two databases.
> Here is a query as the example:
>=20
> select distinct area from ht_booking.hotels;
>=20
> The result from the database on Linux FC3 has no problem but=20
> the one on Windows XP. The data output from the later have=20
> the duplicated rows. It seems to make no difference with or=20
> without "distinct" key word. BUT, this happens only under one=20
> condition: the data in that column contains chinese characters.=20
>=20
> Please feel free to question me if there is anything I  can=20
> do to help with.
> Thank you.

Whare are your database encoding and locale in the two different
installations?

//Magnus

"philip" <philipguo@yahoo.com> writes:
> I installed PostgreSQL 8.0.3 database system on both platforms MS Windows XP
> (sp1) and Linux Fedora 3, and both databse settings are exactly the same
> with the encoding set to 'UNICODE'.

Windows' support for UNICODE is kind of broken I'm afraid :-(.  You
should use a different encoding on that platform, at least until we
figure out an end-run around the OS' limitations.

            regards, tom lane