Re: Writeable CTEs and empty relations

On Thu, Feb 11, 2010 at 12:35 PM, Marko Tiikkaja
<marko(dot)tiikkaja(at)cs(dot)helsinki(dot)fi> wrote:
> On Thu, 11 Feb 2010 19:28:28 +0200, I wrote:
>> On Thu, 11 Feb 2010 10:53:22 -0500, Robert Haas <robertmhaas(at)gmail(dot)com>
>> wrote:
>>> On Thu, Feb 11, 2010 at 8:46 AM, Marko Tiikkaja
>>> <marko(dot)tiikkaja(at)cs(dot)helsinki(dot)fi> wrote:
>>>> On 2010-02-11 03:44 +0200, I wrote:
>>>>> I'm going to have to disappoint a bunch of people and give up. :-(
>>>>
>>>> Btw. would it make sense to apply the WITH-on-top-of-DML part of this
>>>> patch?  At least to me, this seems useful because you can write a
>>>> RECURSIVE SELECT and then use UPDATE .. FROM or DELETE .. USING on that
>>>> CTE.
>>>
>>> Hmm, that's a thought.  Can you split out just that part?
>>
>> Here's the patch.  It's the same as the stuff in writeable CTE patches,
> but
>> I added regression tests.
>
> Whoops.  The reference section in docs still had some traces of writeable
> CTEs.  Updated patch attached.

This looks simple and useful. I haven't tested it, but if it's really
this easy, we should definitely do it.

...Robert