Thread: features required for SQL 92 conformance

features required for SQL 92 conformance

From
"uma chingunde"
Date:
Hi all,
I would like to know if anyone has documented the feature set lacking in
postgresql for conformance with the SQL 92 standard and how these
features can be implemented.
I found information detailing the conformance with SQL 99 but not SQL
92.
Thanks,
Uma

_________________________________________________________________
INDIA TODAY @ Rs. 5 for 5 years !
http://www.indiatoday.com/itoday/intlsubscription/itsubs/it_offer.html
Subcribe Now ...


Re: features required for SQL 92 conformance

From
Josh Berkus
Date:
Uma,

> I would like to know if anyone has documented the feature set lacking in
> postgresql for conformance with the SQL 92 standard and how these
> features can be implemented.
> I found information detailing the conformance with SQL 99 but not SQL
> 92.

Hmmm ... I'm not sure we're missing anything for SQL 92.   Except maybe CREATE
PROCEDURE.

--
-Josh Berkus
 Aglio Database Solutions
 San Francisco


Re: features required for SQL 92 conformance

From
Troels Arvin
Date:
On Thu, 26 Feb 2004 12:32:05 -0800, Josh Berkus wrote:

> I'm not sure we're missing anything for SQL 92.

I'm thinking: Updatable views. As far as I know, SQL-92 has updatable
views in the sense that views deriving from only one base table should be
updatable (as long as the view construction makes it possible for the DBMS
to deduct where to make changes in the base table). Sure, PostgreSQL has
rules, but that hardly qualifies as real support for updatable views.

--
Greetings from Troels Arvin, Copenhagen, Denmark


Re: features required for SQL 92 conformance

From
Josh Berkus
Date:
Troels,

> I'm thinking: Updatable views. As far as I know, SQL-92 has updatable
> views in the sense that views deriving from only one base table should be
> updatable (as long as the view construction makes it possible for the DBMS
> to deduct where to make changes in the base table). Sure, PostgreSQL has
> rules, but that hardly qualifies as real support for updatable views.

Damn, yeah, you're right.    Neil was going to code something so that "WITH
UPDATE" would automatically translate into the creation of 2-3 rules.    But
he got distracted ...

--
-Josh Berkus
 Aglio Database Solutions
 San Francisco


Re: features required for SQL 92 conformance

From
Neil Conway
Date:
Josh Berkus <josh@agliodbs.com> writes:
> Damn, yeah, you're right.  Neil was going to code something so that
> "WITH UPDATE" would automatically translate into the creation of 2-3
> rules.  But he got distracted ...

Indeed, the tragedy of being unable to work on PG full time :-)

But it hardly matters to the subject at hand: as I mentioned on IRC,
AFAIK there is *plenty* of stuff in SQL92 we don't support.

-Neil



Re: features required for SQL 92 conformance

From
Christopher Browne
Date:
neilc@samurai.com (Neil Conway) wrote:
> Josh Berkus <josh@agliodbs.com> writes:
>> Damn, yeah, you're right.  Neil was going to code something so that
>> "WITH UPDATE" would automatically translate into the creation of 2-3
>> rules.  But he got distracted ...
>
> Indeed, the tragedy of being unable to work on PG full time :-)
>
> But it hardly matters to the subject at hand: as I mentioned on IRC,
> AFAIK there is *plenty* of stuff in SQL92 we don't support.

SQL 92 has multiple "levels," and I don't think anyone has gotten
faintly close to supporting every aspect of the higher levels.

- "Entry level" is rudimentary enough that systems that barely feign
  SQL compliance can often comply with it;

- "Transitional level" and "Intermediate level" add in a pretty wide
  set of features, _most_ of which are things PostgreSQL supports;

- "Full SQL92" has features that definitely aren't widely supported,
  although there certainly are some supported by PostgreSQL.

PostgreSQL probably sits somewhere between Transitional and
Intermediate levels, but in some ways, it's a bit irrelevant, as even
if it covered all of particular levels, the NIST organization that
used to be responsible for validating claims of standards conformance
was disbanded back in the '90s.

Claims of "conformance" are a bit specious when there isn't much of a
standards body on this anymore.  And vendors that consider themselves
commercially important are quite prepared to ignore standards whenever
it seems convenient.
--
select 'cbbrowne' || '@' || 'cbbrowne.com';
http://cbbrowne.com/info/advocacy.html
"I withdraw  my claim  that rpm is  proprietary -- my  objections were
based on the documentation for the  version of rpm (2.2.6) that I used
as a  documentation source when  writing makepkg and xrpm."
-- david parsons

Re: features required for SQL 92 conformance

From
Josh Berkus
Date:
Chris,

> Claims of "conformance" are a bit specious when there isn't much of a
> standards body on this anymore.  And vendors that consider themselves
> commercially important are quite prepared to ignore standards whenever
> it seems convenient.

Yeah, why do you think they disbanded the compliance team in the first place?
Just ask Joe Celko ....

--
-Josh Berkus
 Aglio Database Solutions
 San Francisco


Re: features required for SQL 92 conformance

From
Dan Langille
Date:
On Thu, 26 Feb 2004, Josh Berkus wrote:

> Chris,
>
> > Claims of "conformance" are a bit specious when there isn't much of a
> > standards body on this anymore.  And vendors that consider themselves
> > commercially important are quite prepared to ignore standards whenever
> > it seems convenient.
>
> Yeah, why do you think they disbanded the compliance team in the first place?
> Just ask Joe Celko ....

I haven't spoken with Joe in years.. why don't you tell us?

--
Dan Langille - BSDCan: http://www.bsdcan.org/

Re: features required for SQL 92 conformance

From
Josh Berkus
Date:
Dan,

> > > Claims of "conformance" are a bit specious when there isn't much of a
> > > standards body on this anymore.  And vendors that consider themselves
> > > commercially important are quite prepared to ignore standards whenever
> > > it seems convenient.
> >
> > Yeah, why do you think they disbanded the compliance team in the first
> > place? Just ask Joe Celko ....
>
> I haven't spoken with Joe in years.. why don't you tell us?

I thought it was self-evident from my statement.    The largest vendors
weren't happy with their scores on SQL compliance, and by the late 90's had
come to dominate the SQL committee.   So they eliminated conformance testing
so that Oracle, SQL Server, etc. wouldn't look so bad.

And Joe resigned the committee ... probably over that and other things.

--
Josh Berkus
Aglio Database Solutions
San Francisco

Re: features required for SQL 92 conformance

From
Dan Langille
Date:
On Fri, 27 Feb 2004, Josh Berkus wrote:

> Dan,
>
> > > > Claims of "conformance" are a bit specious when there isn't much of a
> > > > standards body on this anymore.  And vendors that consider themselves
> > > > commercially important are quite prepared to ignore standards whenever
> > > > it seems convenient.
> > >
> > > Yeah, why do you think they disbanded the compliance team in the first
> > > place? Just ask Joe Celko ....
> >
> > I haven't spoken with Joe in years.. why don't you tell us?
>
> I thought it was self-evident from my statement.

It wasn't.  That's why I asked.  :)  I'm sure I wasn't the only one.

>    The largest vendors weren't happy with their scores on SQL
> compliance, and by the late 90's had come to dominate the SQL committee.
> So they eliminated conformance testing so that Oracle, SQL Server, etc.
> wouldn't look so bad.

Ahhh, this I did not know.

> And Joe resigned the committee ... probably over that and other things.

Thank you.

--
Dan Langille - BSDCan: http://www.bsdcan.org/

Re: features required for SQL 92 conformance

From
Troels Arvin
Date:
On Thu, 26 Feb 2004 21:13:36 -0500, Neil Conway wrote:

>> Damn, yeah, you're right.  Neil was going to code something so that
>> "WITH UPDATE" would automatically translate into the creation of 2-3
>> rules.  But he got distracted ...
[...]
> But it hardly matters to the subject at hand: as I mentioned on IRC,
> AFAIK there is *plenty* of stuff in SQL92 we don't support.

SQL-92 or not: The lack of updatable views is still an unfortunate
shortcoming.

--
Greetings from Troels Arvin, Copenhagen, Denmark