Date: Tue, 23 Nov 2010 16:19:19 +0100
From: Louis-David Mitterrand <vindex+lists-pgsql-sql@apartia.org>
To: pgsql-sql@postgresql.org
Subject: "compressing" consecutive values into one
Message-ID: <20101123151918.GA22094@apartia.fr>
Mail-Followup-To: pgsql-sql@postgresql.org
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.5.21 (2010-09-15)

Hi,

On time series price data I'm trying to remove consecutive identical
prices and keep only the latest. I tried:

	delete from price where id_price in (select t.id_price2 from (select
			first_value(p.id_price) over w as id_price1,
			nth_value(p.id_price, 2) over w as id_price2,
			first_value(p.price) over w as price1,
			nth_value(p.price,2) over w as price2
			from price p
			window w as (partition by p.id_rate,p.id_cabin_category,p.id_cruise
			order by p.id_price desc rows between unbounded preceding and
			unbounded following)) as t where price1 = price2);

and it mostly works but I have to do several runs to completely
eliminate identical consecutive prices.

Is there a better, one-pass, way?

Thanks,