Re: IN question

From: "Meredith L(dot) Patterson" <mlp(at)thesmartpolitenerd(dot)com>
To: Steve Atkins <steve(at)blighty(dot)com>
Cc: SF PostgreSQL <sfpug(at)postgresql(dot)org>
Subject: Re: IN question
Date: 2008-12-10 22:34:02
Message-ID: 494043DA.3070601@thesmartpolitenerd.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: sfpug

Steve Atkins wrote:
>
> On Dec 10, 2008, at 2:08 PM, A. Elein Mustain wrote:
>
>> On Wed, Dec 10, 2008 at 01:41:01PM -0800, Josh Berkus wrote:
>>> Steve,
>>>
>>>> I'm not so sure there's such a thing as a limit that's too big.
>>>
>>> Sure there is. out-of-memory error.
>>>
>>> Actually, I'd like to see the limit set at work_mem.
>>>
>>> --Josh
>>>
>>
>> I write big, long queries everyday. I would prefer the
>> default be no limit but out of memory. If you must add
>> a limit (why????) then it should NOT be the default.
>
> Well, one reason for a limit is to provide the DBA with a
> last line of defense against idiot clients. Given some of
> the dumb things automated query builders and ORMs
> are prone to do that's not such a bad idea.

Back in the pre-7.0 days, there was a limit of something like 16384
bytes, which I can see being a problem. But work_mem defaults to 1MB and
is often much larger. How large are the queries these automated query
builders produce? IO/network bottlenecks anyone? I don't care if you're
doing it over dedicated fiber, if you're passing a query larger than 1MB
you're doing it wrong.

--mlp

In response to

Responses

Browse sfpug by date

  From Date Subject
Next Message Jeff Davis 2008-12-10 23:39:33 Re: IN question
Previous Message Steve Atkins 2008-12-10 22:12:10 Re: IN question