I think you are on something here.
Below is what concerns me most in the example:
data[o.name].append(o.model3_set.all().order_by('created')[0])
This will sort the entire model3 table just to get the min value.
And if the db is doing this 6000 times, that is a bit useless.
Have you tried using raw SQL in you model?
See: http://docs.djangoproject.com/en/dev/topics/db/sql/#topics-db-sql
You do not have to use the Django ORM for this.
Maeck
On Dec 8, 7:43 am, DavidA <[EMAIL PROTECTED]> wrote:
> If I undestand the problem correctly, in MySQL you could do this in
> one query as:
>
> select
> m.*,
> (select min(created) from model2 where id = m.model2_id) as
> first_created,
> (select max(created) from model2 where id = m.model2_id) as
> last_created
> from model1 m
> ;
>
> I don't know how that translates to the Django ORM (or even if it
> does), though.
>
> -Dave
>
> On Dec 8, 12:27 am, Malcolm Tredinnick <[EMAIL PROTECTED]>
> wrote:
>
> > On Sat, 2008-12-06 at 20:56 -0800, erikcw wrote:
> > > Hi all,
>
> > > I'm trying to write a model query that will return a queryset along
> > > with the latest (and earliest) data from 2 related models.
>
> > > Right now I'm doing something like this:
>
> > > objects = Model1.objects.filter(user=3).select_related() #about 6,000
> > > objects
>
> > > data = {}
> > > for o in objects:
> > > data[o.name] = [o.field1, o.field2]
> > > data[o.name].append(o.field3.model2_set.all().latest('created'))
> > > #get latest row from related model2
> > > data[o.name].append(o.model3_set.all().order_by('created')[0])
> > > #get earliest row from related model3
>
> > > The problem is that this results in a TON of database queries. This
> > > view is taking over a minute to process. The select_related on the
> > > first line doesn't seem to be helping since I'm using latest()/
> > > order_by which generates a new query.
>
> > > How can I make this more efficient? Denormalizing the isn't an option
> > > since model2 and model 3 are many-to-one.
>
> > By the way (since I still haven't worked out the complex SQL to make it
> > three queries yet), this last statement isn't correct. Denormalising
> > doesn't just mean flattening. You can store a computed dependent field
> > in model1 for the related information. So you could, for example, store
> > the relevant date and pk value of the related model2 row and update that
> > whenever a new model2 is inserted.
>
> > Malcolm
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Django users" group.
To post to this group, send email to django-users@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/django-users?hl=en
-~----------~----~----~----~------~----~------~--~---
