Thread: Updateable views was:(Re: [HACKERS] Proposal for SYNONYMS)

Updateable views was:(Re: [HACKERS] Proposal for SYNONYMS)

From
"Jaime Casanova"
Date:
On 3/9/06, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Josh Berkus <josh@agliodbs.com> writes:
> > Eh?  I thought that it was just syntatic sugar that was missing.   I've
> > built lots of updatable views manually; I don't see what's difficult about
> > it.
>
> I think you'll find that corner cases like inserts involving nextval()
> don't work real well with a rule-based updatable view.  But perhaps I'm
> just scarred by the many complaints we've had about rules.  With a plain
> unconditional DO INSTEAD rule it might be OK ...
>
>                        regards, tom lane
>

the last time i talk with Bernd Helmle, he was preparing the code to
send to patches for discussion... that was two months ago...

the current code had problems with casts and i think with domains too...

i will contact with Bernd to know if he did some more work, if not i
can send to patches the latest path he sent me...

--
regards,
Jaime Casanova

"What they (MySQL) lose in usability, they gain back in benchmarks, and that's
all that matters: getting the wrong answer really fast."                          Randal L. Schwartz


Re: Updateable views was:(Re: [HACKERS] Proposal for SYNONYMS)

From
Richard Huxton
Date:
Jaime Casanova wrote:
> On 3/9/06, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> Josh Berkus <josh@agliodbs.com> writes:
>>> Eh?  I thought that it was just syntatic sugar that was missing.   I've
>>> built lots of updatable views manually; I don't see what's difficult about
>>> it.
>> I think you'll find that corner cases like inserts involving nextval()
>> don't work real well with a rule-based updatable view.  But perhaps I'm
>> just scarred by the many complaints we've had about rules.  With a plain
>> unconditional DO INSTEAD rule it might be OK ...

> the last time i talk with Bernd Helmle, he was preparing the code to
> send to patches for discussion... that was two months ago...
> 
> the current code had problems with casts and i think with domains too...
> 
> i will contact with Bernd to know if he did some more work, if not i
> can send to patches the latest path he sent me...

I'd certainly be interested in having auto-updatable views in 8.2 - even 
if it was only for the simplest of cases. If I can be of any help 
testing etc. let me know.

--   Richard Huxton  Archonet Ltd


Re: Updateable views was:(Re: [HACKERS] Proposal for

From
Bernd Helmle
Date:

--On Freitag, März 10, 2006 09:43:04 +0000 Richard Huxton
<dev@archonet.com> wrote:

> I'd certainly be interested in having auto-updatable views in 8.2 - even
> if it was only for the simplest of cases. If I can be of any help testing
> etc. let me know.


Yeah, that would be cool. I've sent the latest patch to -hackers. Feel free
to check it out. I currently know that array fields (e.g. field[1]) causes
problems, but i'm pretty sure there's much more work left...
       Bernd