Date: Tue, 15 Jan 2008 14:20:45 +0100
From: =?UTF-8?B?TWFnbmUgTcOmaHJl?= <Magne.Mahre@Sun.COM>
Subject: Re: JOIN a table twice for different values in the same query
In-reply-to: <47869EF7.70106@denterprises.org>
To: Colin Wetherbee <cww@denterprises.org>
Cc: Phillip Smith <phillip.smith@weatherbeeta.com.au>,
 pgsql-sql@postgresql.org
Message-id: <478CB32D.6050304@sun.com>
Organization: Sun Microsystems Inc.
MIME-version: 1.0
Content-type: text/plain; format=flowed; charset=UTF-8
Content-transfer-encoding: 7BIT
References: <002f01c853d7$63bc0680$9b0014ac@wbaus090>
	<47869C8C.4080109@denterprises.org> <47869EF7.70106@denterprises.org>
User-Agent: Thunderbird 2.0.0.9 (X11/20071119)

Colin Wetherbee wrote:
> Colin Wetherbee wrote:
>> Phillip Smith wrote:
>>> As a side note - all the IATA codes are unique for each airport -
>>> wouldn't it be better to use these as the Primary Key and Foreign
>>> Keys? Then you wouldn't have to even join the tables unless you
>>> wanted the port names (not just the code)
>>
>> This is true, but FWIW, my application will mostly be joining for the 
>> name of the airport or the city, not the code.
>>
>> I'll keep the idea of using the codes as keys in mind, though.  Thanks 
>> for pointing that out.
> 
> Oh, now I remember why I'm using IDs as keys. ;)
> 
> The code isn't always going to be an airport, and, for example, a train 
> station in Buenos Aires could conceivably have the same code as a 
> shipping port in Rotterdam, which, in turn, might well be JFK. :)

Note that IATA codes are _NOT_ unique.   The current list of IATA 
trigrams list upward of 300 duplicate codes.  If you include the train 
stations, there might be additional collisions.

You could consider using the ICAO four-letter identifiers instead. They 
are unique, and are preferred by airspace management authorities.  A 
mapping to the corresponding IATA code exists.

--Magne