DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com;
	h=Message-ID:Received:Date:From:Subject:To:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding;
	b=hhl2/8qZi5dQI1WSV1j0GROnJBQssN3tZHn1Y4H2zWLLgf9KNvv4bjksavE0FWxig+V4e2Oyj9z9URVrxT8leByLciz5SrBiPoQVCWoMsJVTKAcrmovj+QyJWl869Ih/aKBjcZoSZk+/ai7HHrjlKWEWniVzutEW9K1xIbUdKj0=
	;
Message-ID: <20060606185639.4573.qmail@web33309.mail.mud.yahoo.com>
Date: Tue, 6 Jun 2006 11:56:39 -0700 (PDT)
From: <operationsengineer1@yahoo.com>
Subject: Re: How To Exclude True Values
To: Richard Broersma Jr <rabroersma@yahoo.com>,
	"pgsql-sql@postgresql.org" <pgsql-sql@postgresql.org>
In-Reply-To: <20060606173628.17134.qmail@web31815.mail.mud.yahoo.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit

> > inspect_id, inspect_result_id,
> inspect_result_pass,
> > inspect_result_timestamp
> > 3, 5, f, 2006-06-05 05:00:00
> > 3, 6, t, 2006-06-05 06:00:00
> > 4, 7, f, 2006-06-05 07:00:00
> > *4, 8, f, 2006-06-05 08:00:00*
> > the query linked in this post will return 3
> lines...
> > 
> > 1, 2, f
> > 3, 5, f
> > *4, 8, f* -- the only one i really want
> 
> 
> From your sample it seems to me that you are really
> only looking for the most recient occuring
> record that have produced a false test regardless of
> which Inspect_id or inspect_result_id it came
> from.  Is this correct?

Richard, not quite.  i have an inspection entry form
and a defect entry form.  

an inspection node (each row in t_inspect is an
inspection node) that has passed can't have a new
defect added - since it has already passed.  

therefore, in the defect entry form, i only want to
display those inspection nodes that don't have a true
value.  by definition, a true value will always be the
last (by time) inspect_result.

therefore, i need all the inspect nodes that don't yet
have a true value (iow, a true value in the last (by
time) inspect_result_pass row).

an inspection node can have multiple inspection
results, hence, the t_inspection_results table.

this might seem counter-intuitive at first, but it
makes sense since it may take 5 tries to eventually
pass a particular inspection node (f, f, f, f, t) for
fucntional test, for example.  one node, five tests to
pass it.

i went to this model b/c a unit can actually pass
through the various test nodes multiple times.  a unit
might go through the same QA node three times - and i
need to differentiate between these three QA nodes.

if each inspection node only occurred once, i could
use one table and link it a t_inspect_node table.  due
to the multiple passes, though, i can't do this in
this case.

i hope this makes sense.  i'm living it and it is
sometimes hard to keep straight.

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com