In-Reply-To: <20070427001004.B5F5D9FBABA@postgresql.org>
References: <20070427001004.B5F5D9FBABA@postgresql.org>
Mime-Version: 1.0 (Apple Message framework v752.3)
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
Message-Id: <D9565B27-E46B-498D-ACCE-6DA953DA1ADA@earlham.edu>
Cc: PostgreSQL SQL List <pgsql-sql@postgresql.org>
Content-Transfer-Encoding: 7bit
From: Kevin Hunter <hunteke@earlham.edu>
Subject: Re: Selecting rows with "static" ordering
Date: Fri, 27 Apr 2007 10:11:56 -0400
To: Steve Midgley <public@misuse.org>

At 8:01p -0400 on 26 Apr 2007, Steve Midgley wrote:
> From that application, I want to retrieve all those rows, and I  
> want them in the order they are currently stored in that variable.  
> So take for example this foreign application variable:
>
>   ids = "3,2,5,1,4"
>
> The application then executes this sql:
>
>   select * from table where id in (3,2,5,1,4)

> Is there a "neat trick" that anyone knows for pulling this off in a  
> single query? Basically right now I'm issuing 5 queries to the  
> backend to ensure ordering but this horribly inefficient.

Why not do this particular ordering on the application side?  With  
this example SQL, and assuming you've PRIMARY KEYed( id ), you're  
guaranteed to get a unique id for each tuple.  Presumably, you've got  
the order you want stored within your application, so just pull it  
out of a hash in the order you need it.

Or is it terribly more complicated than this?

Kevin