This is gonna be django performance series
The Best Way to Learn is to Teach
I'm a man of few words
So let's start with example of Users and Cities models.
# models.py
class City(models.Model):
name = models.CharField(max_length=63)
class User(models.Model):
fullname = models.CharField(max_length=127)
city = models.ForeignKey(City, on_delete=models.SET_NULL, null=True)
And this is our serializers
# serializers.py
class CitySerializer(serializers.ModelSerializer):
class Meta:
model = City
fields = ('id', 'name')
class UserSerializer(serializers.ModelSerializer):
city = CitySerializer()
class Meta:
model = User
fields = ('id', 'fullname', 'city')
And this is our views
# views.py
class UserListView(ListAPIView):
serializer_class = UserSerializer
queryset = User.objects.all()
So if you looked to the queryset User.objects.all()
its sql is
SELECT "users"."id","users"."fullname","users"."city_id"
FROM "users"
It is not efficient because for each user you are going to access database to select user city for the nested serializer CitySerializer
, It is (n+1) queries
Solution: use select_related
after changes queryset will be User.objects.select_related('city')
sql:
SELECT "user"."id","user"."fullname", "user"."city_id","city"."id",
"city"."name" FROM "user"
LEFT OUTER JOIN "city" ON ("user"."city_id" = "city"."id")
So now it is just one query
Top comments (2)
Nice tip Man
Nice tip :D