Add a field to raw data to prevent serialization error - django

Question

Assume I have two models and their serializers like this:
class Billing(models.Model):
...
class Transaction(models.Model):
billing = models.ForeignKey(Billing, null=False, blank=False)
...
class TransactionSerializer(serializers.ModelSerializer):
billing = serializers.PrimaryKeyRelatedField(queryset=Billing.objects.all())
class Meta:
model = Transaction
fields = '__all__'
Now I want to have an endpoint to post new transaction to a billing, something like this:
post http://address/billings/{id}/transactions [{other fields except billing because the billing exists in the address}]
For this purpose I have written a viewset like this:
class BillingTransactionList(generics.ListCreateAPIView):
serializer_class = TransactionSerializer
def get_queryset(self):
billing = get_object_or_404(Billing.objects.all(), pk=self.kwargs['pk'])
return Transation.objects.filter(billing=billing)
def perform_create(self, serializer):
billing = get_object_or_404(Billing.objects.all(), pk=self.kwargs['pk'])
return serializer.save(billing=billing)
But if billing does not exist in data that I get from request, serializer would be failed becuase it needs the billing in raw data of request. I have the billing from endpoint and I just want that serializer accepts the data and further I will add the billing as I have done in perform_create.
There is an option to add required=False to TransactionSerializer but I need this serializer in another places with required=True, Also there is an another solution to write another serializer, but in my real example the serializer is a big class and I do not want to write it again. I am looking for a simple solution to just ignore the disappearance of billing in data and let me define it whenever I want.
I am using django 1.11.3 and DRF 3.8.2.

Answer 1

I think it's better to use 2 different serializers (in a way that they have a parent serializer with most of the common code in it) for your actions.
but if you don't want to go with the 2 serializers solution, you can override the to_internal_value method and pick the url parameter(pk) for billing field if it is empty in the raw data (also, since the generic views pass themselves to serializers you have access to the url parameter in the serializer). so:
class TransactionSerializer(serializers.ModelSerializer):
billing = serializers.PrimaryKeyRelatedField(queryset=Billing.objects.all())
class Meta:
model = Transaction
fields = '__all__'
def to_internal_value(self, data):
billing_pk_in_url = self.context['view'].kwargs.get('pk', None)
if 'billing' not in data: ## or any other condition that you want
data['billing'] = billing_pk_in_url
return super().to_internal_value(data)
and now, you don't even have to override the perform_create method:
class BillingTransactionList(generics.ListCreateAPIView):
serializer_class = TransactionSerializer
def get_queryset(self):
billing = get_object_or_404(Billing.objects.all(), pk=self.kwargs['pk'])
return Transation.objects.filter(billing=billing)

Answer 2

Can you change your design? I think that would be easier. For example since billing is a field on Transaction, you could query http://address/billings/transactions?billing_id={billing_id}
Attempting to get a ViewSet to support detail operation when the instance doesn't exist seems like a difficult thing to do. It would be better to approach it from another way that allows for the billing to be nullable from the start.
If that's not an option you need to stop using get_object_or_404 if it's possible that the id passed in the url doesn't exist. You also should change the TransactionSerializer as such:
class TransactionSerializer(serializers.ModelSerializer):
billing = serializers.PrimaryKeyRelatedField(
queryset=Billing.objects.all(), allow_null=True, many=False)
...