Date: Sun, 13 Nov 2005 12:24:01 -0300
From: Alvaro Herrera <alvherre@commandprompt.com>
To: "Goulet, Dick" <DGoulet@vicr.com>
Cc: Frank van Vugt <ftm.van.vugt@foxi.nl>, Greg Stark <gsstark@mit.edu>,
	pgsql-interfaces@postgresql.org
Subject: Re: Incremental results from libpq
Message-ID: <20051113152401.GB31570@surnet.cl>
Mail-Followup-To: "Goulet, Dick" <DGoulet@vicr.com>,
	Frank van Vugt <ftm.van.vugt@foxi.nl>, Greg Stark <gsstark@mit.edu>,
	pgsql-interfaces@postgresql.org
References: 
 <4001DEAF7DF9BD498B58B45051FBEA6502EF5389@25exch1.vicorpower.vicr.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: 
 <4001DEAF7DF9BD498B58B45051FBEA6502EF5389@25exch1.vicorpower.vicr.com>
User-Agent: Mutt/1.5.10i

Goulet, Dick wrote:

> 	For a simple "select * from <table_name>" even with a where
> clause you may simply get the results one row at a time as they are
> extracted & deemed appropriate for the result set.  But if you've
> included a group by or order by clause, or a union/intersect/minus
> clause then yes, the database will assemble results before sending them
> to the client.

So, what happens with the 1/x query Tom mentioned?  How does Oracle
handles that situation?  Notice there's no special clause in the query
itself, so if it's extracted and returned, there's no way for the server
to know that there's a problem laying ahead.

> 	Also Oracle supports cursors which allow you to pop one or more
> rows off of the result set at a time.  The way Postgresql returns data
> all at once is different. 

Postgres supports cursors too.  The Qt guys, and everyone else, could be
using it to get incremental results right now, no libpq mods necessary.

-- 
Alvaro Herrera                                http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.