Re: 7.4 To Do

From: Paul Ramsey <pramsey(at)refractions(dot)net>
To: Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
Cc: PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: 7.4 To Do
Date: 2002-12-01 04:45:28
Message-ID: 3DE993E8.4EF881AE@refractions.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

We recently finished some evaluations of Oracle, and one of the things
which jumped out at me, PostgreSQL booster that I am, was that in 9i
Oracle has finally surpassed PostgreSQL in some elements of
object-relational technology. Among the things you can do are:

- Create new compound object types on the SQL command line. So for
example, the "imaginary numbers" example in the PostgreSLQ manual could
all be done on the command line.
- Address components of objects using dot-notation. (select
employee.salary from employees)
- Index objects based on their components (create index blah_idx on
employees (employee.last_name))

In combination with ARRAY data types, and references, you can do pretty
fancy things without ever creating a C library.

PostgreSQL seems to have most of the underpinning already. I even did
some experimenting with the "tables as a datatype" stuff. However, it's
definately incomplete. In the tables-as-datatype example, you cannot get
the objects to reconstruct themselves, you have to do it by hand. And
while oid's make handy reference holders, there does not seem to be any
particular performance benefit to doing object-oriented modelling over
relational modelling in PostgreSQL at this point. By contrast, OORDBMS
like Informix can be blazingly fast when used with the proper OO model.
(Oracle performance enhancement for OO models is mixed at best right
now.)

Bruce Momjian wrote:

> Does anyone have additions for the list. I didn't see anything
> discussed that jumped out at me, though I didn't see "Conquer the world"
> on there. ;-)

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2002-12-01 04:53:33 Re: Hard-coded PUBLIC in pg_dump
Previous Message Tom Lane 2002-12-01 04:28:47 Re: Wishlist for 7.4: Plan stability