Envelope-to: pgsql-sql@postgresql.org
Message-ID: <4BDBEEC7.5030405@bankir.ru>
Date: Sat, 01 May 2010 13:05:11 +0400
From: silly sad <sad@bankir.ru>
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.5)
	Gecko/20100111 Thunderbird/3.0
MIME-Version: 1.0
To: pgsql-sql@postgresql.org
Subject: Re: LIMIT problem
References: <4BDAD20A.1090503@bankir.ru> <hrfqqn$u2r$1@reversiblemaps.ath.cx>
In-Reply-To: <hrfqqn$u2r$1@reversiblemaps.ath.cx>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

On 05/01/10 03:58, Jasen Betts wrote:
> On 2010-04-30, silly sad<sad@bankir.ru>  wrote:
>> suppose i request
>>
>> SELECT foo(t.x) FROM t LIMIT 1;
>>
>> Whither it DEFINED how many times foo() will be executed?
>
> foo will be executed repeatedly until it returns a result or all the
> rows in t are exhausted.
>
>> May anyone rely on it?
>
> not sure
>
>> Or we have to avoid this non SQLish trick?
>
>
> This will execute it once (or not at all where t has no rows)
>
>   SELECT foo(x) FROM (SELECT x FROM t LIMIT 1) as bar;
>

this subselect "isolation" looks like a safe way.
Unlike LIMITed select looks unpredictable to me.

Thanx for assistance :)