DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma;
	h=mime-version:in-reply-to:references:date:message-id:subject:from:to
	:cc:content-type:content-transfer-encoding;
	b=iO8vGbuINuAPboWy2m9n+VhzW44+HiqkThOkNRvKtJxl81z//6GbLYbghjG+KI9ZKu
	xU9T3MtOpVNo67W3tQoiZJSyw40iTPTGIzpOjXyz/ySfb8DpIlimrlgSeAlCgp7QKi8j
	o/kv+n6JvMvlYQRPyN8MCcptsyHYu5rduM4/c=
MIME-Version: 1.0
In-Reply-To: <C629C7E8.35E64%maximilian.tyrtania@onlinehome.de>
References: <dcc563d10905071154j613847edqc5cab8c8af68d496@mail.gmail.com>
	<C629C7E8.35E64%maximilian.tyrtania@onlinehome.de>
Date: Fri, 8 May 2009 05:35:21 -0600
Message-ID: <dcc563d10905080435y48690c3cvddce77a5dcf7c1ae@mail.gmail.com>
Subject: Re: Distinct oddity
From: Scott Marlowe <scott.marlowe@gmail.com>
To: Maximilian Tyrtania <maximilian.tyrtania@onlinehome.de>
Cc: pgsql-sql@postgresql.org
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

On Fri, May 8, 2009 at 3:28 AM, Maximilian Tyrtania
<maximilian.tyrtania@onlinehome.de> wrote:
> am 07.05.2009 20:54 Uhr schrieb Scott Marlowe unter scott.marlowe@gmail.c=
om:
>
>> On Thu, May 7, 2009 at 2:21 AM, Maximilian Tyrtania
>> <maximilian.tyrtania@onlinehome.de> wrote:
>>> Hi there,
>>>
>>> does this look right?
>>>
>>> FAKDB=3D# select count(distinct(f.land)) from firmen f where
>>> f.typlist=3D'Redaktion';
>>> =A0count
>>> -------
>>> =A01975
>>> (1 row)
>>>
>>> FAKDB=3D# select count(distinct(f.land||'1')) from firmen f where
>>> f.typlist=3D'Redaktion';
>>> =A0count
>>> -------
>>> =A04944
>>> (1 row)
>>
>> Yeah, that does seem odd. =A0Could it be something like nulls in your
>> data set? =A0just guessing really. =A0If you could make a small test cas=
e
>> that shows it happening and allows others to reproduce it you're
>> likely to get more bites.
>
> It doesn't seem to be related to null values (which wouldn't explain it
> anyway) nor to this particular field...
>
> FAKDB=3D# select count(*) from firmen where bezeichnung is null;
> =A0count
> -------
> =A0 =A0 0
> (1 row)

That's not the same field as in the original query.


> My attempts at reproducing this with a freshly created table failed, of
> course.

Instead of trying to create a test case from scratch, isolate some
rows that cause this, put them in another table, and then pg_dump that
one table, cleaned as needed for privacy, here.