I have models like this

class Survey(models.Model):
    name = models.CharField(max_length=120)
    ...


class Question(models.Model):
    survey = models.ForeignKey(Survey,
                               related_name='questions',
                               on_delete=models.CASCADE)
    ...


class Option(models.Model):
    question = models.ForeignKey(Question,
                                 related_name='options',
                                 on_delete=models.CASCADE)
    ...

Serializers looks like:

class OptionSerializer(serializers.ModelSerializer):
    class Meta:
        model = Option
        exclude = ('question','id')


class QuestionSerializer(serializers.ModelSerializer):
    options = OptionSerializer(many=True)
    class Meta:
        model = Question
        exclude = ('survey',)


class SurveyDetailSerializer(serializers.ModelSerializer):
    creator = serializers.StringRelatedField(read_only=True)
    questions = QuestionSerializer(many=True)

    class Meta:
        model = Survey
        exclude = ('is_active',)

GET request is clear, i get survey with all questions and options

views.py

class SurveyDetailView(generics.RetrieveAPIView):
    """Detail Survey view with all survey's questions"""
    queryset = Survey.objects.all()
    serializer_class = SurveyDetailSerializer

Result

{
    "id": 3,
    "creator": "nvbr",
    "questions": [
        {
            "id": 5,
            "options": [
                {
                    "text": "OPTION A"
                },
                {
                    "text": "OPTION B"
                }
            ],
            "answer_type": "single option",
            "text": "QUESTION 1?"
        }
    ],
    "name": "stackoverflow survey",
    "description": "Example",
    "created_at": "2021-09-23T13:07:19.679174Z"
}

but how can I make POST to save Survey, Questions and Options? Or should I always create separate endpoints for all models? SPA frontend if it important

CodePudding user response：

Ah, welcome to the limitations of DRF.

Presently, the way to work with nested data is by overriding the update method on your serializer:

serializers.py

class SurveyDetailSerializer(serializers.ModelSerializer):

    ...

    def update(self, instance, validated_data):

        # get the questions array:
        questions = validated_data.pop('questions')
        
        # create new questions, map to survey 
        ...

        # continue with regular update method:
        super().update(self, instance, validated_data)