Django ORM & Django REST Framework: How to make "grouped" arrays?

Question

Currently, my Django server can return the JSON below:

[
  {
    "id": 1,
    "customer": {
      "id": 1,
      "name": "John Doe"
    },
    "description1": "...",
    "description2": "...",
    "description3": "...",
    "description4": "..."
  },
  {
    "id": 2,
    "customer": {
      "id": 1,
      "name": "John Doe"
    },
    "description1": "...",
    "description2": "...",
    "description3": "...",
    "description4": "..."
  },
  {
    "id": 3,
    "customer": {
      "id": 1,
      "name": "John Doe"
    },
    "description1": "...",
    "description2": "...",
    "description3": "...",
    "description4": "..."
  },
  {
    "id": 4,
    "customer": {
      "id": 2,
      "name": "Jane Doe"
    },
    "description1": "...",
    "description2": "...",
    "description3": "...",
    "description4": "..."
  },
  {
    "id": 5,
    "customer": {
      "id": 2,
      "name": "Jane Doe"
    },
    "description1": "...",
    "description2": "...",
    "description3": "...",
    "description4": "..."
  }
]

My models are:

from django.db import models

class Customer(models.Model):
    name = models.CharField(verbose_name="Customer name")

from django.db import models


class PurchaseDescriptions(models.Model):
    customer =  models.ManyToManyField("customer.Customer", related_name="customer", verbose_name="Customer")
    description1 = models.CharField(verbose_name="Description 1")
    description2 = models.CharField(verbose_name="Description 2")
    description3 = models.CharField(verbose_name="Description 3")
    description4 = models.CharField(verbose_name="Description 4")

my serializers are:

from rest_framework import serializers

class CustomerSerializer(serializers.ModelSerializer):
    class Meta:
        model = Customer
        fields = [
            "id",
            "name",
        ]

class PurchaseDescriptionsSerializer(serializers.ModelSerializer):
    customer = CustomerSerializer()
    
    class Meta:
        model = PurchaseDescriptions
        fields = [
            "id",
            "customer",
            "description1",
            "description2",
            "description3",
            "description4",
        ]

and my viewset is:

from rest_framework.viewsets import GenericViewSet
from rest_framework import mixins

class ScheduleViewSet(
    mixins.ListModelMixin,
    GenericViewSet,
):
    queryset = PurchaseDescriptions.objects.all()
    serializer_class = PurchaseDescriptionsSerializer

I want to rewrite it and make some groups based on customer, e.g.

[
  {
    "customer": {
      "id": 1,
      "name": "John Doe"
    },
    "data": [
      {
        "id": 1,
        "description1": "...",
        "description2": "...",
        "description3": "...",
        "description4": "..."
      },
      {
        "id": 2,
        "description1": "...",
        "description2": "...",
        "description3": "...",
        "description4": "..."
      },
      {
        "id": 3,
        "description1": "...",
        "description2": "...",
        "description3": "...",
        "description4": "..."
      }
    ]
  },
  {
    "customer": {
      "id": 2,
      "name": "Jane Doe"
    },
    "data": [
      {
        "id": 4,
        "description1": "...",
        "description2": "...",
        "description3": "...",
        "description4": "..."
      },
      {
        "id": 5,
        "description1": "...",
        "description2": "...",
        "description3": "...",
        "description4": "..."
      }
    ]
  }
]

But I don’t know how to make this using Django’s views and serializers. How can I get such JSON? Note that I want to avoid creating additional tables in the database if it is possible.

Asked By: Nicnem

||

Source

Answer 1

You can override the serializers to_representation() to get custom JSON response.

Give this a try:

class PurchaseDescriptionsSerializer(serializers.ModelSerializer):
    customer = CustomerSerializer()

    class Meta:
        model = PurchaseDescriptions
        fields = [
            "id",
            "customer",
            "description1",
            "description2",
            "description3",
            "description4",
        ]

    def to_representation(self, instance):
        data = super().to_representation(instance)
        customer_id = data["customer"]["id"]
        customer_name = data["customer"]["name"]
        del data["customer"]
        return {
            "customer": {
                "id": customer_id,
                "name": customer_name,
            },
            "data": [data],
        }

Answered By: Sumithran

Answer 2

# defined model 

class Customer(models.Model):
    name = models.CharField(verbose_name="Customer name")

class PurchaseDescriptions(models.Model):
    customer =  models.ManyToManyField("Customer", related_name="customer", verbose_name="Customer")
    description1 = models.CharField(verbose_name="Description 1")
    description2 = models.CharField(verbose_name="Description 2")
    description3 = models.CharField(verbose_name="Description 3")
    description4 = models.CharField(verbose_name="Description 4")

    
   

# serializers.py
from rest_framework.serializers import ModelSerializer, Serializer
from rest_framework.fields import CharField, SerializerMethodField
from drf_spectacular.utils import extend_schema_field


class PurchaseDescriptionSchema(ModelSerializer):
    class Meta:
        model = PurchaseDescriptions
        exclude = ("customer",)


class PurchaseDescriptionsSchema(Serializer):
    customer = SerializerMethodField(method_name="nested_customer")
    data = PurchaseDescriptionSchema(many=True)
    
    # i don't like MethodField 
    # but there is no other way because it's too wierd schema structure
    @extend_schema_field(OpenApiTypes.OBJECT)
    def nested_customer(self, obj: Customer2) -> Dict[str,Any]:
        return {
            "id": obj.id,
            "name": obj.name
        }
    


# viewSets.py
class ScheduleViewSet(
    mixins.ListModelMixin,
    GenericViewSet,
):
    queryset = Customer.objects.prefetch_related(
        Prefetch(lookup="purchasedescriptions_set", to_attr="data", queryset=PurchaseDescriptions.objects.all())
    ).all()
    serializer_class = PurchaseDescriptionsSchema

ViewSetList’s response

즐코하세요

++++

If you want to implement it in a way that doesn’t use a MethodField, you can use a Concat in a QuerySet to get the customer data in JSON format.

Customer.objects.annotate(customer=Concat(Value("{ "id": "), id, Value(","name" "), name, Value("}") )).prefetch_related(~~~)

You can configure your queryset this way and use the

class PurchaseDescriptionsSchema(Serializer):
    # set customer = SerializerMethodField(method_name="nested_customer")
    customer = JsonField() # DRF SerializerField
    data = PurchaseDescriptionSchema(many=True)

implement the Serializer as shown above.

but, above case’s queryset is too messy, so I use SerializerMethodField.

Answered By: SoungRyoul Kim

Django ORM & Django REST Framework: How to make "grouped" arrays?

Question:

Answers: