如何将OrderedDict转换为Django模型实例？附报错与代码示例

现有Django模型 #

class City(models.Model):
    name = models.CharField(max_length=50, default='')

class Street(models.Model):
    name = models.CharField(max_length=50)
    city = models.ForeignKey(City, on_delete=models.CASCADE)

class Shop(models.Model):
    name = models.CharField(max_length=50)
    city = models.ForeignKey(City, on_delete=models.SET_NULL, null=True)
    street = models.ForeignKey(Street, on_delete=models.SET_NULL, null=True)
    house = models.SmallIntegerField()
    opening_time = models.TimeField()
    closing_time = models.TimeField()

对应序列化器代码 #

from rest_framework import serializers

class CitySerializer(serializers.ModelSerializer):
    class Meta:
        model = City
        fields = ['id', 'name']


class StreetSerializer(serializers.ModelSerializer):
    city = CitySerializer()

    class Meta:
        model =  Street
        fields = ['id', 'name', 'city']

    # 原方法无法工作，问题核心在此
    def create(self, validated_data):
        city_data = validated_data.pop('city')
        city = City.objects.create(**city_data)
        street = Street.objects.create(**validated_data)


class ShopSerializer(serializers.ModelSerializer):
    city = CitySerializer()
    street = StreetSerializer()

    class Meta:
        model = Shop
        fields = ['id', 'name', 'city', 'street', 'house', 'opening_time', 'closing_time']


    def create(self, validated_data):
        city_data = validated_data.pop('city')
        city = City.objects.create(name=city_data['name'])
        street_data = validated_data.pop('street')
        street = Street.objects.create(name = street_data['name'], city = city)
        shop = Shop.objects.create(**validated_data)
        # 此处报错：NoneType，因为ForeignKey不能用add()
        shop.city.add(city)
        # 赋值后未保存
        shop.street = street
        return shop 

运行时错误信息 #

ValueError: Cannot assign "OrderedDict([('name', 'Ryazan')])": "Street.city" must be a "City" instance.

问题分析与修复方案 #

1. 修复StreetSerializer的create方法 #

错误根源是StreetSerializer嵌套了CitySerializer，导致validated_data['city']是OrderedDict而非City实例，必须先将字典转换为模型实例再关联：

class StreetSerializer(serializers.ModelSerializer):
    city = CitySerializer()

    class Meta:
        model =  Street
        fields = ['id', 'name', 'city']

    def create(self, validated_data):
        # 取出城市的字典数据
        city_data = validated_data.pop('city')
        # 用get_or_create避免重复创建同名城市
        city, _ = City.objects.get_or_create(**city_data)
        # 将City实例传入Street创建逻辑
        street = Street.objects.create(city=city, **validated_data)
        return street

2. 修复ShopSerializer的create方法 #

原代码存在两个问题：

• shop.city.add(city)错误：ForeignKey是一对一关联，没有add()方法，直接赋值即可
• 赋值shop.street = street后未保存实例

优化后的代码：

class ShopSerializer(serializers.ModelSerializer):
    city = CitySerializer()
    street = StreetSerializer()

    class Meta:
        model = Shop
        fields = ['id', 'name', 'city', 'street', 'house', 'opening_time', 'closing_time']

    def create(self, validated_data):
        # 处理城市数据，避免重复创建
        city_data = validated_data.pop('city')
        city, _ = City.objects.get_or_create(name=city_data['name'])
        
        # 处理街道数据，关联已创建的城市
        street_data = validated_data.pop('street')
        street, _ = Street.objects.get_or_create(name=street_data['name'], city=city)
        
        # 创建Shop时直接传入关联实例，一步到位
        shop = Shop.objects.create(city=city, street=street, **validated_data)
        return shop

简化方案（可选） #

如果不需要每次创建新的City/Street，而是关联已有数据，建议用SlugRelatedField替代嵌套序列化器，直接通过名称关联实例，无需手动转换字典：

class StreetSerializer(serializers.ModelSerializer):
    city = serializers.SlugRelatedField(slug_field='name', queryset=City.objects.all())

    class Meta:
        model = Street
        fields = ['id', 'name', 'city']

此时传入{"name": "某街道", "city": "Ryazan"}即可自动关联到对应的City实例，省去字典转模型的步骤。

内容的提问来源于stack exchange，提问作者mezigar