Reputation: 17
I am trying to optimize a Django project (vers. 1.8.6) in which each page shows 100 companies and their data at once. I noticed that an unnecessary amount of SQL queries (especially with contact.get_order_count
) are performed within the index.html snippet below:
index.html:
{% for company in company_list %}
<tr>
<td>{{ company.name }}</td>
<td>{{ company.get_order_count }}</td>
<td>{{ company.get_order_sum|floatformat:2 }}</td>
<td><input type="checkbox" name="select{{company.pk}}" id=""></td>
</tr>
{% for contact in company.contacts.all %}
<tr>
<td> </td>
<td>{{ contact.first_name }} {{ contact.last_name }}</td>
<td>Orders: {{ contact.get_order_count }}</td>
<td></td>
</tr>
{% endfor %}
{% endfor %}
The problem seems to lie in constant SQL queries to other tables using foreign keys. I looked up how to solve this and found out that prefetch_related()
seems to be the solution. However, I keep getting a TemplateSyntaxError about being unable the parse the prefetch, no matter what parameter I use. What is the proper prefetch syntax, or is there any other way to optimize this that I missed?
I've included relevant snippets of model.py below in case it's relevant. I got prefetch_related
to work in the defined methods, but it doesn't change the performance or query amount.
model.py:
class Company(models.Model):
name = models.CharField(max_length=150)
def get_order_count(self):
return self.orders.count()
def get_order_sum(self):
return self.orders.aggregate(Sum('total'))['total__sum']
class Contact(models.Model):
company = models.ForeignKey(
Company, related_name="contacts", on_delete=models.PROTECT)
first_name = models.CharField(max_length=150)
last_name = models.CharField(max_length=150, blank=True)
def get_order_count(self):
return self.orders.count()
class Order(models.Model):
company = models.ForeignKey(Company, related_name="orders")
contact = models.ForeignKey(Contact, related_name="orders")
total = models.DecimalField(max_digits=18, decimal_places=9)
def __str__(self):
return "%s" % self.order_number
EDIT:
The view is a ListView and defines the company_list
as model = Company
. I altered the view based on given suggestions:
class IndexView(ListView):
template_name = "mailer/index.html"
model = Company
contacts = Contact.objects.annotate(order_count=Count('orders'))
contact_list = Company.objects.all().prefetch_related(Prefetch('contacts', queryset=contacts))
paginate_by = 100
Upvotes: 0
Views: 1320
Reputation: 309039
Calling the get_order_count
and get_order_sum
methods causes one query every time the method is called. You can avoid this by annotating the queryset.
from django.db.models import Count, Sum
contacts = Contact.objects.annotate(order_count=Count('orders'), order_sum=Sum('orders'))
You then need to use a Prefetch
object to tell Django to use your annotated queryset.
contact_list = Company.objects.all().prefetch_related(Prefetch("contacts", queryset=contacts)
Note that you need to add the prefetch_related
to your queryset in the view, it is not possible to call it in the template.
Since you are using ListView
, you should be overriding the get_queryset
method, and calling prefetch_related()
there:
class IndexView(ListView):
template_name = "mailer/index.html"
model = Company
paginate_by = 100
def get_queryset(self):
# Annotate the contacts with the order counts and sums
contacts = Contact.objects.annotate(order_count=Count('orders')
queryset = super(IndexView, self).get_queryset()
# Annotate the companies with order_count and order_sum
queryset = queryset.annotate(order_count=Count('orders'), order_sum=Sum('orders'))
# Prefetch the related contacts. Use the annotated queryset from before
queryset = queryset.prefetch_related(Prefetch('contacts', queryset=contacts))
return queryset
Then in your template, you should use {{ contact.order_count }}
instead of {{ contact.get_order_count }}
, and {{ company.order_count }}
instead of {{ company.get_order_count }}
.
Upvotes: 1
Reputation: 168
Try this in views.py
company_list = Company.objects.all().prefetch_related("order", "contacts")
Upvotes: 0