Django Rest框架：在序列化程序中预取初始数据

Question

I know that django has select_related and prefetch_related that can be used when querying items from database to increase its performance, and it can be used in pair with django rest framework's nested serializer.

However, the problem come when I want to use the serializer to create my model for example:

class CompanySerializer(serializer.serializers):
    employee_set = serializers.JSONField()
    class Meta:
        model = Company
        fields = ('id', 'employee_set')
    def create(self, validated_data):
        employee_set = validated_data.pop('employee_set')
        for employee in employee_set:
            serializer = EmployeeSerializer(data=employee)
            serializer.is_valid(raise_exception=True)
            serializer.save()

class EmployeeSerializer(serializer.serializers):
    card = serializers.PrimaryKeyRelatedField(queryset=Card.objects.all())
    class Meta:
        model = Employee
        fields = ('id', 'name', 'card')
    def validate(self, obj):
        if card.employee_set.all().count() > 3:
            raise serializers.ValidationError({'_error': 'invalid})
        return data

For example, I want to create a company with multiple employee like:

request.POST:

{
    employee_set: [
        { name: 'tim', card: 1 },
        { name: 'bob', card: 1 },
        { name: 'jimmy', card: 2},
    ]
}

then I can use CompanySerializer(request.POST), right?

However, when I am saving this serializer, the EmployeeSerializer will iterate over each employee and query employee.card_set, which result in a lot of sql queries. Is there any way to do it similar like prefetch_related?

thanks

Answer 1

As the question is too open, I'll give the directions.

Django ORM allows you to create several instances at once using bulk_create .

Also note that you should not call another serializer from a serializer's create/update, as shown in the documentation

Answer 2

You can send extra information into your serializer, via context

# use select_related or prefetech_related to get this value
card_counters = {'card_id1': 1, 'card_id2': 3, ...}

for employee in employee_set:
    serializer = EmployeeSerializer(
        data=employee, 
        context={'card_counter': card_counters.get(employee['card_id'], 0)}
    )
    ...

And then use it when doing the validation:

def validate(self, obj):
    if self.context['card_counter'] > 3:
        raise serializers.ValidationError({'_error': 'invalid})
    return data