On Fri, Feb 27, 2009 at 5:49 PM, saxon75 <saxo...@gmail.com> wrote:
>
> I'm having a somewhat odd error occurring when I try to use
> select_related() and annotate() in the same queryset.
>
> The application I'm working on is a blog-type app where posts can be a
> variety of types--article, book review, film review, etc. I use multi-
> table inheritance with a single base class, Node, and then each post
> type with its own subclass, e.g. Article, BookReview, etc. Each Node
> has an FK relationship with an Archetype, which essentially just
> contains a string saying whether the Node references an Article,
> BookReview, or whatever. Each Node also has an FK relationship with a
> Section (used for grouping) and with several Comments.
>
> Most of my views do some relatively simple queries and then pass a
> queryset of Nodes to a base template, which then iterates over the
> queryset and calls a custom template tag in each iteration to
> determine what kind of node it is and render it.
>
> In the view I'm currently working on, I use the
> list_detail.object_list generic view with a wrapper around it. In
> order to speed things up, I added select_related() to my queryset,
> like so:
>
> Node.objects.select_related().filter(section__keyword=sect,
> published=True).order_by('-created')
>
> This worked pretty well, reducing the number of times the DB is hit in
> the template significantly.
>
> Now, something else I do in the template is display the number of
> Comments associated with each Node. That causes a couple of extra DB
> hits per template iteration. So I thought that using the annotate()
> function would work nicely. Like so:
>
> Node.objects.select_related().filter(section__keyword=sect,
> published=True).order_by('-created').annotate(comment_count=Count
> ('comment'))
>
> Except, when I do that, something gets all messed up with the
> resulting queryset. Instances in the queryset have data in the wrong
> fields. This becomes apparent when I try to use my custom template
> tag:
>
> def render_node(node, user):
> template_name = node.archetype.classname.lower() + '_detail.html'
> edit = False
> if user.is_authenticated():
> if user.is_superuser or user.groups.filter
> (name=node.section.keyword).count():
> edit = True
> return render_to_string(template_name, {'node': node, 'user': user,
> 'edit': edit})
>
> This throws an exception, saying that "'long' object has no attribute
> 'lower'." When I put in some diagnostic code to print
> node.archetype.classname to the console, it comes out as a number,
> which makes no sense. Then when I tried printing node.comment_count,
> it returns "Article," which is what should be returned by
> node.archetype.classname.
>
> If I remove select_related(), it works like I expect it should, except
> that now the db is hit a bunch of extra times each time through the
> loop. If I remove annotate(), it also works.
>
> Has anyone else seen behavior like this?
>
> For reference, I am using r9781.
> >
>
If possible I would start by SVN upping, r9808 seems to indicate it fixes
percisely this problem.
Alex
--
"I disapprove of what you say, but I will defend to the death your right to
say it." --Voltaire
"The people's good is the highest law."--Cicero
--~--~---------~--~----~------------~-------~--~----~
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
django-users+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/django-users?hl=en
-~----------~----~----~----~------~----~------~--~---
