Received: from ute.cnri.reston.va.us (cnri170.cnri.reston.va.us
	[132.151.1.170])
	by postgresql.org (8.11.3/8.11.1) with ESMTP id f32K73K39913
	for <pgsql-interfaces@postgresql.org>;
	Mon, 2 Apr 2001 16:07:03 -0400 (EDT)
	(envelope-from akuchlin@mems-exchange.org)
Received: from akuchlin by ute.cnri.reston.va.us with local (Exim 3.20 #1)
	id 14kAbM-0000uW-00
	for pgsql-interfaces@postgresql.org; Mon, 02 Apr 2001 16:07:00 -0400
From: Andrew Kuchling <akuchlin@mems-exchange.org>
To: pgsql-interfaces@postgresql.org
Subject: pgdb.py fetching same row multiple times
Reply-to: akuchlin@mems-exchange.org
Message-Id: <E14kAbM-0000uW-00@ute.cnri.reston.va.us>
Date: Mon, 02 Apr 2001 16:07:00 -0400
X-Archive-Number: 200104/3
X-Sequence-Number: 1386

With the 7.1RC1 code, plus two unrelated fixes to pgdb.py that are in
the CVS, the .fetch() method of pgdbCursor always returns the first
row, not each successive row.  For example, this test program:

import pgdb
db = pgdb.connect('ute.cnri.reston.va.us:mems:akuchlin')

c = db.cursor()
c.execute('select * from fab_providers')
print c.fetchone()
print c.fetchone()
db.close()

will print the same row twice:

ute akuchlin>python /tmp/t.py
['CNF', 'Cornell Nanofabrication Facility', ... ]
['CNF', 'Cornell Nanofabrication Facility', ... ]
ute akuchlin>

The first column is a primary key, so the table doesn't actually
contain two duplicate rows.  

Has anyone else noted this?  Anyone have a fix?

(Note that this doesn't apply to the basic interface implemented by
pgmodule.c; the problem is only with the pgdb.py wrapper to provide a
DB-API interface.)

--amk