I was just looking at all the upcoming features scheduled to make it
into 8.3, and with all those goodies, wouldn't it make sense for this
to be a 9.0 release instead of an 8.3? It looks like postgresql is
rapidly catching up to oracle if 8.3 branch gets every feature
scheduled for it.

About the only big features pg 8.3 doesn't have is materialized views
and RMAN..

Now that PostgreSQL is getting so close to oracle functionality, is
there any worry in the community that oracle will begin to target
postgres like they're targeting mySQL?

regards,
karen

On 29 Jan 2007 13:25:31 -0800, Karen Hill <karen_hill22@yahoo.com> wrote:
> I was just looking at all the upcoming features scheduled to make it
> into 8.3, and with all those goodies, wouldn't it make sense for this
> to be a 9.0 release instead of an 8.3? It looks like postgresql is
> rapidly catching up to oracle if 8.3 branch gets every feature
> scheduled for it.

Well I see it in two ways. For one, the features are certainly
great and a significant advance. This alone could mandate version
bump to 9.0.

On the other hand, the 8.x line is so successful I would like it to
stay for a copule revisions more. Well, it does have a nice feeling
about it: "What? Yeah, it does support windowing function, we've
introduced them around version 8.3. Naah, no big deal, wait for
the version 8.4, you'll be surprosed. Naah, we keep version 9.0
for truly significant changes". And I must say, I do like it.

> About the only big features pg 8.3 doesn't have is materialized views
> and RMAN..

Personally I'm missing two things, which were discussed in the
past, but would be nice to have:
* more efficient storage of varlen data -- some time ago there were
ideas to get rid of constant 4-bytes for length and use more elastic
approach. Smaller tables, bigger performance.
* updatable views [ or am I missing something? ] -- it seems to me
they were close to be completed, but I don't remember if they were
completed and committed or not.

Regards,
Dawid

---------------------------(end of broadcast)---------------------------
TIP 1: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to majordomo@postgresql.org so that your
message can get through to the mailing list cleanly

> Personally I'm missing two things, which were discussed in the
> past, but would be nice to have:
> * more efficient storage of varlen data -- some time ago there were
> ideas to get rid of constant 4-bytes for length and use more elastic
> approach. Smaller tables, bigger performance.
> * updatable views [ or am I missing something? ] -- it seems to me
> they were close to be completed, but I don't remember if they were
> completed and committed or not.


I'm missing stuff like true polymorphic function arguments and return
values (where I can mix different datatypes and do variable number of
parameters), also I personally hate 'select * from my_func() as table(x
varchar)' syntax... system should be able to omit the table structure
definition and pick it up from function return.

Oh well, back to work.

Where do we submit wishlist entries, anyway?

Peter

"Dawid Kuroczko" <qnex42@gmail.com> writes:

> * updatable views [ or am I missing something? ] -- it seems to me
> they were close to be completed, but I don't remember if they were
> completed and committed or not.

Something different than rules?
(http://www.postgresql.org/docs/8.2/i...ive/rules.html) (They exist for a
while, I've just linked the latest released docs...)

--
Jorge Godoy <jgodoy@gmail.com>

---------------------------(end of broadcast)---------------------------
TIP 6: explain analyze is your friend

Jorge Godoy wrote:
> "Dawid Kuroczko" <qnex42@gmail.com> writes:
>
>
>> * updatable views [ or am I missing something? ] -- it seems to me
>> they were close to be completed, but I don't remember if they were
>> completed and committed or not.
>>
>
> Something different than rules?
> (http://www.postgresql.org/docs/8.2/i...ive/rules.html) (They exist for a
> while, I've just linked the latest released docs...)
>
Quite. Rules are not updateable views. Rules are a hacked up way to
create an updateable view. The patch
as discussed IIRC, would make the rules automatically.


Sincerely,

Joshua D. Drake




---------------------------(end of broadcast)---------------------------
TIP 4: Have you searched our list archives?

http://archives.postgresql.org/

Jorge Godoy wrote:
> "Dawid Kuroczko" <qnex42@gmail.com> writes:
> > * updatable views [ or am I missing something? ] -- it seems to me
> > they were close to be completed, but I don't remember if they were
> > completed and committed or not.
>
> Something different than rules?
> (http://www.postgresql.org/docs/8.2/i...ive/rules.html) (They
> exist for a while, I've just linked the latest released docs...)

See http://developer.postgresql.org/inde...pdatable_views for
further wisdom.

--
Peter Eisentraut
http://developer.postgresql.org/~petere/

---------------------------(end of broadcast)---------------------------
TIP 1: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to majordomo@postgresql.org so that your
message can get through to the mailing list cleanly

On Tue, 2007-01-30 at 02:35 -0800, Joshua D. Drake wrote:
> > Something different than rules?
> > (http://www.postgresql.org/docs/8.2/i...ive/rules.html) (They exist for a
> > while, I've just linked the latest released docs...)
> >
> Quite. Rules are not updateable views. Rules are a hacked up way to
> create an updateable view.

I wouldn't go that far. Rules can do things that updatable views can't
do. Sometimes a view can't be updatable because an update to that view
would be ambiguous (as far as the system knows), but you can still use
the well-defined rules system to *tell* the system what you want an
update to mean.

Updatable views provide a subset of the functionality of rules, but they
do it automatically without much effort on the part of the DBA. That's
great, but it won't replace rules.

Regards,
Jeff Davis


---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match

On Jan 29, 11:06 pm, qne...@gmail.com ("Dawid Kuroczko") wrote:

> * updatable views [ or am I missing something? ] -- it seems to me
> they were close to be completed, but I don't remember if they were
> completed and committed or not.
>

PostgreSQL has updatable views via the rules system. I use updatable
views all the time in postgres.

On 30 Jan 2007 12:15:17 -0800, Karen Hill <karen_hill22@yahoo.com> wrote:
> On Jan 29, 11:06 pm, qne...@gmail.com ("Dawid Kuroczko") wrote:
> > * updatable views [ or am I missing something? ] -- it seems to me
> > they were close to be completed, but I don't remember if they were
> > completed and committed or not.
>
> PostgreSQL has updatable views via the rules system. I use updatable
> views all the time in postgres.

That is not a point really. This todo is not about implementing rule
system which PostgreSQL already has.

It's about implementing infrastructure to set up updatable views automatically,
as the standard dictates. And this is a feaure PostgreSQL lacks. If you
want updatable views you have to issue couple of CREATE RULEs apart
from CREATE VIEW. The point is that you could create updatable views
with sole CREATE VIEW command.

Another example is table partitioning which PostgreSQL has and doesn't
have. You can set up table partitioning with clever set of triggers and
table inheritance, but it lacks explicit DDLs to do so.

Regards,
Dawid

---------------------------(end of broadcast)---------------------------
TIP 5: don't forget to increase your free space map settings

Jeff Davis <pgsql@j-davis.com> writes:
> Updatable views provide a subset of the functionality of rules, but they
> do it automatically without much effort on the part of the DBA. That's
> great, but it won't replace rules.

Exactly --- but there is also a place for a low-effort, "do the right
thing" feature, which is pretty much what the SQL spec's concept of
updatable views is.

There's also some question about whether our current definition of rules
can even support a non-surprising implementation of an updatable view.
The issue of multiple evaluation of what might be volatile expressions
keeps coming up...

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 2: Don't 'kill -9' the postmaster