Re: [HACKERS] Well, then you keep your darn columns

From: Hannu Krosing <hannu(at)tm(dot)ee>
To: The Hermit Hacker <scrappy(at)hub(dot)org>
Cc: Adriaan Joubert <a(dot)joubert(at)albourne(dot)com>, Peter Eisentraut <peter_e(at)gmx(dot)net>, pgsql-hackers(at)postgreSQL(dot)org
Subject: Re: [HACKERS] Well, then you keep your darn columns
Date: 2000-01-24 21:52:07
Message-ID: 388CC987.B5B86454@tm.ee
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

The Hermit Hacker wrote:
>
> On Mon, 24 Jan 2000, Adriaan Joubert wrote:
>
> > As a user let me just say that I have been waiting for all the
> > 'ALTER TABLE..' commands for a long time (changing constraints would be great
> > too!) and it has been mentioned many times that one should not use oids in
> > application code. The way I do it at the moment is to copy the table, so I
> > think all users are used to losing their oids when remocing a column from a
> > table. So while it is certainly better to keep the oids, Peter's code is a
> > big improvement on the current situation.
>
> Except, as Chris Bitmead brought up, OIDs appear to be a key requirement
> in ODBMSs ... so, if we want to go what I *think* is 'next generation',
> OIDs have to be kept ...

But the decision was (from Vadim IIRC) to drop them, at least in non system
tables.
The cited reasons were:
* crappy implementation that taxed performance (probably fixed by now)
* nobody else seemed to have them and the push then was to the direction of
mainstream bean-counting DB with main objective of getting that base
functionality right.
* they take up "too much" space (probably a non-issue in current world of
dropping disk/memory prices)
* you can always re-implement them at the application level (the same was
cited
for dropping time travel)

If it is now a general opinion that OIDs will remain, it should be stated
somewhere,
as I suspect that much of pg community operates under the impression that they
are
going away in future as have some other nice but not fully developed features
like
time travel.

I do understand that time travel could be implemented using a bunch of
rules+views
but to be really useful rules views should be much more developed and
integrated
with inheritance, up to the level where you could say:

create table stock(
item_id int,
quantity int,
price numeric(15,2)
) inherits (time_travel);

and then have all the needed tables/rules/views produced automatically.

If we had that level of sofistication we could also safely drop OIDs as a
built-in
system feature and still have it by changing the last line to

) inherits (time_travel,odbms_table);

-----------------
Hannu

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Oliver Elphick 2000-01-24 21:52:56 Inheritance, referential integrity and other constraints
Previous Message Ansley, Michael 2000-01-24 21:39:27 RE: [HACKERS] Well, then you keep your darn columns