Message-ID: <01BE8E3F.959B0140.jbarnett@pobox.com>
From: Jon Barnett <jbarnett@pobox.com>
Reply-To: "jbarnett@pobox.com" <jbarnett@pobox.com>
To: "'herouth maoz'" <herouth@oumail.openu.ac.il>,
	"'pgsql-interfaces@hub.org'" <pgsql-interfaces@hub.org>
Subject: RE: [INTERFACES] JDBC next() method
Date: Sat, 24 Apr 1999 10:45:34 +1000
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit

On Saturday, 24 April 1999 6:45, herouth maoz [SMTP:herouth@oumail.openu.ac.il] 
wrote:
> No, the question is whether Postgres's behaviour - of returning a row with
> a null field in case no rows fell into the WHERE category - is compatible
> with other databases, or indeed with the SQL standard. I don't have the
> standard in front of me at this moment. But you must understand that this
> is how PostgreSQL behaves, no matter whether your frontend is in Java,
> python, perl or C. It's not a question of JDBC.

I concede your point on that.  I'm just obsessing over the ultimate result of 
this quirk(?) - independence of the application.

From a practicalities point of view, if the result is not compliant with the 
SQL standard, the question is whether to hide it in the abstraction layer (JDBC 
driver) or fix it at the source - I guess costs and complexity of the fix will 
determine that (although it is always better to fix it at the source).  Of 
course, hiding it in the abstraction layer means you need to do the same for 
every interface (which is a headache for maintainers).

This is just speculation and musing, in any case, as I am not sure what the SQL 
standard says with respect to this limit case (where a table is empty).  More 
research needed I suppose. :)

JonB.