Re: proposal: set GUC variables for single query - Mailing list pgsql-hackers

From Jan Urbański
Subject Re: proposal: set GUC variables for single query
Date
Msg-id 4E9BEC83.5070603@wulczer.org
Whole thread Raw
In response to Re: proposal: set GUC variables for single query  (Robert Haas <robertmhaas@gmail.com>)
Responses Re: proposal: set GUC variables for single query
List pgsql-hackers
On 17/10/11 02:53, Robert Haas wrote:
> On Sun, Oct 16, 2011 at 4:58 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> Dimitri Fontaine <dimitri@2ndQuadrant.fr> writes:
>>> Now that you mention it, the following might actually already work:
>>
>>>  WITH settings AS (
>>>    SELECT set_config('timezone', 'Europe/Amsterdam', t),
>>>           set_config('work_mem', '1 GB', t)
>>>  ),
>>>       foo AS (
>>>    SELECT …
>>>  )
>>>  INSERT INTO bar SELECT * FROM foo;
>>
>> Only for small values of "work" ... you won't be able to affect planner
>> settings that way, nor can you assume that that WITH item is executed
>> before all else.  See recent thread pointing out that setting values
>> mid-query is unsafe.
> 
> I previously floated the idea of using a new keyword, possibly LET,
> for this, like this:
> 
> LET var = value [, ...] IN query

LET was something I thought about, although you'd have to use something
like parenthesis around the GUC assignements because "value" can contain
commas, leading to shift/reduce conflicts (that sucks, unfortunately).

But before whipping out the paint bucket I wanted to see if there's
enough buy-in to justify rehashing the syntax details.

Cheers,
Jan


pgsql-hackers by date:

Previous
From: Yeb Havinga
Date:
Subject: Re: [REVIEW] Patch for cursor calling with named parameters
Next
From: Florian Pflug
Date:
Subject: Re: Underspecified window queries in regression tests