Message-ID: <48771344.3020305@sime.com>
Date: Fri, 11 Jul 2008 10:01:08 +0200
From: Mario Weilguni <mweilguni@sime.com>
User-Agent: Thunderbird 2.0.0.14 (X11/20080505)
MIME-Version: 1.0
To: Ow Mun Heng <Ow.Mun.Heng@wdc.com>
CC: Alvaro Herrera <alvherre@commandprompt.com>,
	Pgsql performance <pgsql-performance@postgresql.org>
Subject: Re: Altering a column type - Most efficient way
References: <1215676380.15319.2.camel@neuromancer.home.net>	
	<4875C9FA.7020101@sime.com>	
	<1215682481.15319.14.camel@neuromancer.home.net>	
	<20080710135756.GD3757@alvh.no-ip.org>
	<1215742620.853.10.camel@neuromancer.home.net>
In-Reply-To: <1215742620.853.10.camel@neuromancer.home.net>
Content-Type: text/plain; charset=ISO-8859-15; format=flowed
Content-Transfer-Encoding: 7bit

Ow Mun Heng schrieb:
> On Thu, 2008-07-10 at 09:57 -0400, Alvaro Herrera wrote:
>   
>> Ow Mun Heng wrote:
>>
>>     
>>> This is what I see on the table
>>>
>>> NEW attypmod = -1
>>> OLD attypmod =  8
>>>       
>> 8 means varchar(4) which is what you said you had (4+4)
>> -1 means unlimited size.
>>
>>     
>
> This is cool. 
>
> If it were this simple a change, I'm not certain why (I believe) PG is
> checking each and every row to see if it will fit into the new column
> definition/type. 
>
> Thus, I'm still a bit hesitant to do the change, although it is
> definitely a very enticing thing to do. ( I presume also that this
> change will be instantaneous and does not need to check on each and
> every row of the table?)
>
> Thanks./
>
>   

It should be safe, because the length limit is checked at insert/update 
time, and internally, a varchar(20) is treated as something like this:
foo      varchar(1000000000) check (length(foo) <= 20)


The change is done without re-checking all rows, and will not fail IF 
the new size is longer than the old size.