validating json fields - django-rest-framework

Question

This is my serializer file
from rest_framework import serializers
#All validations here
class BudgetSerializer(serializers.Serializer):
start_date = serializers.DateTimeField(format="%Y-%m-%d %H:%M:%S")
end_date = serializers.DateTimeField(format="%Y-%m-%d %H:%M:%S")
lifetime_budget = serializers.FloatField(max_value=None, min_value=None,allow_null=True)
def validate(self, attrs):
if attrs['start_date'] > attrs['end_date']:
raise serializers.ValidationError("start_date must be less than end date")
return attrs
class IoSerializer(serializers.Serializer):
name = serializers.CharField()
state = serializers.CharField()
currency = serializers.CharField(read_only=True)
budget_type = serializers.CharField(read_only=True)
budget_intervals = BudgetSerializer(many=True)
Json which i am serializing has 1000's of fields but i only want to validate few fields and return all json fields.But problem with serialzer is that it only outputs json consisting of mentioned field in serialiser.
Like i am sending json with 1000s of object names but it only return json which consist of start_date,end_date,lifetime_budget,name because it is mentioned in serializer file.I wanted to only validate few fields so i wrote them in serializer file but only this fields are returned.
My views.py code
class InsertionOrderViewset(viewsets.ViewSet):
def create(self, request, format=None):
advertiser_id = request.query_params.get('advertiser_id', None)
data = json.dumps(request.data)
io_object = json.loads(data, object_hook=lambda x: namedtuple('io_object', x.keys())(*x.values()))
serializer = IoSerializer(data=request.data, instance=io_object)
serializer.is_valid(raise_exception=True)
response_data = IO.create(data=request.data, params={"advertiser_id": advertiser_id })
return Response(request.data)

Answer 1

serializers.data method will only return values which are supplied to it unless you declare some read_only fields and or do any of these - supply an instance during serializer initialization, return an instance from serializer.save method. When serializer.instance has a value then read_only fields are returned with serializer.data. Your use case suggests that the incoming JSON object has 1000s of data of which only a small portion requires validation. My suggestion will be not to pass all the data, pluck only the necessary portion of the data, send them to serializer and do validation, return serializer.errors if any else rejoin the serializer.data with the incoming JSON and return the JSON. Since, majority of the data has nothing to do with validation, passing them in to the serializer will cause significant performance issue due to the way drf-serializer validates and maps values.
Still if you still want to supply all of the JSON to serializer and validate only a few consider supplying an instance when initializing serializer, dynamically declare ReadOnlyField in your serializer.__init__
serializers.py
from rest_framework import serializers
# default attributes of objects
default_attrs = ['__doc__', '__module__', '__slots__', '_fields', '_source']
class BudgetSerializer(serializers.Serializer):
start_date = serializers.DateTimeField(format="%Y-%m-%d %H:%M:%S")
end_date = serializers.DateTimeField(format="%Y-%m-%d %H:%M:%S")
lifetime_budget = serializers.FloatField(max_value=None, min_value=None,allow_null=True)
def __init__(self, *args, **kwargs):
super(BudgetSerializer, self).__init__(*args, **kwargs)
instance = kwargs.get('instance', None)
for attr in dir(instance):
if attr not in self.fields and not callable(getattr(instance, attr)) and attr not in default_attrs:
self.fields[attr] = serializers.ReadOnlyField()
def validate(self, attrs):
if attrs['start_date'] > attrs['end_date']:
raise serializers.ValidationError("start_date must be less than end date")
return attrs
class IoSerializer(serializers.Serializer):
name = serializers.CharField()
state = serializers.CharField()
currency = serializers.CharField(read_only=True)
budget_type = serializers.CharField(read_only=True)
budget_intervals = BudgetSerializer(many=True)
def __init__(self, *args, **kwargs):
super(IoSerializer, self).__init__(*args, **kwargs)
instance = kwargs.get('instance', None)
for attr in dir(instance):
if attr not in self.fields and not callable(getattr(instance, attr)) and attr not in default_attrs:
self.fields[attr] = serializers.ReadOnlyField()
inside views.py
import json
from collections import namedtuple
# inside view
data = json.dumps(request.data)
io_object = json.loads(data, object_hook=lamdba x: namedtuple('io_object', x.keys())(*x.values()))
serializer = IoSerializer(data=request.data, instance=io_object)
serializer.is_valid(raise_exception=True)
print(serializer.data)