Composite Primary/Foreign Keys

72 views
Skip to first unread message

slango

unread,
Nov 27, 2009, 3:49:10 PM11/27/09
to Django developers
Hello all. I'm new to the Django community, but I have an itch that
needs scratching and some time off to scratch it. The itch, of course,
is composite primary and foreign keys in Django.

I've started to speak with David Cramer on the issue. Several months
ago, he was pursuing a solution which he has documented here:
http://code.djangoproject.com/wiki/MultipleColumnPrimaryKeys. David no
longer believes that he was taking the right approach, so before I
begin any work on this, I wanted to check the pulse of the developer
community on the matter. Hopefully we can start a discussion to
determine what the right approach is, and I can get to work
implementing it. With a solid plan in place, perhaps others will want
to join the cause too. The goal is to create a solution that will be
worth introducing to the core Django distribution.

So, if you have a moment, let's get things started here. It is
understood that not everyone *needs* composite primary and foreign
keys, but this is something that plenty of people do need, and it is
in the interest of making the Django model layer more complete.

Russell Keith-Magee

unread,
Nov 28, 2009, 2:30:46 AM11/28/09
to django-d...@googlegroups.com
This is a long-standing issue for Django (ticket #373), and yes, we
would like to have a solution. From a brief look at the wiki, it
appears that the known issues and design constraints are well
documented.

However - with that said, we're currently about 1 month out from the
deadline for adding new features to Django 1.2. The features that are
scheduled for that release were chosen back in October, following
several months of discussion and design plans. The core team is
currently focussed on developing the features that we want in 1.2.

As a result, you may find it difficult to get design feedback (or
contributions) from those in the core team. I don't want to discourage
you from working on this - if this is your itch, please do scratch it.
I just want to let you know why you may not get the design feedback
you need or want.

On a positive note, If you want to consider this as a problem for the
1.3 timeframe, working through the issues now means that you will be
able to present a solid design, and possibly even an implementation
when the feature selection phase for 1.3 comes around. Working
implementations always attract positive attention during the early
phases.

Yours,
Russ Magee %-)

Luke Plant

unread,
Nov 28, 2009, 2:27:00 PM11/28/09
to django-d...@googlegroups.com
On Saturday 28 November 2009 07:30:46 Russell Keith-Magee wrote:

> However - with that said, we're currently about 1 month out from
> the deadline for adding new features to Django 1.2. The features
> that are scheduled for that release were chosen back in October,
> following several months of discussion and design plans. The core
> team is currently focussed on developing the features that we want
> in 1.2.

Isn't this in our "Medium priority" list for 1.2 ?
http://code.djangoproject.com/wiki/Version1.2Features

Malcolm is down at a committer for this, but no-one as lieutenant, and
it looks like someone voted -1. As far as I understand it, that means
this is definitely a possibility for 1.2, but would require a lot of
work in terms of a good design and implementation that all the core
devs are happy with.

Regards,

Luke

--
"He knows the way I take: when he has tried me, I shall come forth
as gold" (Job 23:10).

Luke Plant || http://lukeplant.me.uk/

Russell Keith-Magee

unread,
Nov 28, 2009, 7:47:13 PM11/28/09
to django-d...@googlegroups.com
On Sun, Nov 29, 2009 at 3:27 AM, Luke Plant <L.Pla...@cantab.net> wrote:
> On Saturday 28 November 2009 07:30:46 Russell Keith-Magee wrote:
>
>> However - with that said, we're currently about 1 month out from
>>  the deadline for adding new features to Django 1.2. The features
>>  that are scheduled for that release were chosen back in October,
>>  following several months of discussion and design plans. The core
>>  team is currently focussed on developing the features that we want
>>  in 1.2.
>
> Isn't this in our "Medium priority" list for 1.2 ?
> http://code.djangoproject.com/wiki/Version1.2Features
>
> Malcolm is down at a committer for this, but no-one as lieutenant, and
> it looks like someone voted -1. As far as I understand it, that means
> this is definitely a possibility for 1.2, but would require a lot of
> work in terms of a good design and implementation that all the core
> devs are happy with.

Gah - my mistake. I had forgotten that Malcolm had claimed this one.
The -1 vote is James Bennett, and it appears to be a -1 on the fact
that there is no working prototype patch on the ticket.

In practice, I'm not sure this changes much. Given that Malcolm is
still buried in his work commitments, if the current state of the
ticket is "we need to do some design work", one month from the feature
deadline isn't the best time to start those design discussions.

Russ %-)
Reply all
Reply to author
Forward
0 new messages