如何修改Django序列化器输出，按项目分组合同数据？

问题描述 #

我想要修改Django序列化器的输出格式。基于类的视图中有get_projects方法，它返回Contract实例的queryset：

projects = Project.objects.all()
contracts = Contract.objects.filter(project__in=projects)

Contract与Project是一对多关系：

class Contract(models.Model):
    project = models.ForeignKey(Project)

当前使用的MySerializer如下：

class MySerializer(serializers.Serializer):
    project_guid = serializers.UUIDField(source='guid')
    project_name = serializers.CharField(source='name')
    contract_guid = serializers.UUIDField(source='guid')
    contract_number = serializers.CharField(source='number')

该序列化器的输出为：

[
  {
    "projectGuid": "project_guid_1",
    "projectName": "project_name_1",
    "contractGuid": "contract_guid_1",
    "contractNumber": "contract_number_1"
  },
  {
    "projectGuid": "project_guid_1",
    "projectName": "project_name_1",
    "contractGuid": "contract_guid_2",
    "contractNumber": "contract_number_2"
  },
  {
    "projectGuid": "project_guid_2",
    "projectName": "project_name_2",
    "contractGuid": "contract_guid_4",
    "contractNumber": "contract_number_4"
  },
  {
    "projectGuid": "project_guid_2",
    "projectName": "project_name_2",
    "contractGuid": "contract_guid_5",
    "contractNumber": "contract_number_5"
  }
]

希望将输出修改为按项目分组合同的格式，示例如下：

[
 {
  "project_guid": "project_guid_1",
  "project_name": "project_name_1",
  "contracts": [
    {
      "contract_guid": "contract_guid_1",
      "contract_number": "contract_number_1"
    },
    {
      "contract_guid": "contract_guid_2",
      "contract_number": "contract_number_2"
    }
  ]
 },
 {
  "project_guid": "project_guid_2",
  "project_name": "project_name_2",
  "contracts": [
    {
      "contract_guid": "contract_guid_4",
      "contract_number": "contract_number_4"
    },
    {
      "contract_guid": "contract_guid_5",
      "contract_number": "contract_number_5"
    }
  ]
 }
]

请问如何实现这一需求？

解决方案 #

可以通过嵌套序列化器+优化查询逻辑的方式实现，步骤如下：

1. 定义合同子序列化器 #

先创建仅处理合同字段的序列化器，用于嵌套展示：

class ContractSerializer(serializers.Serializer):
    contract_guid = serializers.UUIDField(source='guid')
    contract_number = serializers.CharField(source='number')

2. 定义项目主序列化器 #

创建处理项目信息的序列化器，通过SerializerMethodField获取并序列化当前项目的所有合同：

class ProjectSerializer(serializers.Serializer):
    project_guid = serializers.UUIDField(source='guid')
    project_name = serializers.CharField(source='name')
    contracts = serializers.SerializerMethodField()

    def get_contracts(self, obj):
        # 从项目对象反向关联获取所有合同
        contracts = obj.contract_set.all()
        return ContractSerializer(contracts, many=True).data

3. 调整视图查询逻辑 #

将查询起点从合同改为项目，同时用prefetch_related预加载关联合同，避免N+1查询问题：

# 替换原有的contracts查询，改用预加载合同的项目queryset
projects = Project.objects.all().prefetch_related('contract_set')

4. 使用项目序列化器输出 #

在视图中用ProjectSerializer处理项目queryset，返回序列化后的数据：

serializer = ProjectSerializer(projects, many=True)
return Response(serializer.data)

可选：用ModelSerializer简化代码 #

如果模型字段与输出字段匹配度高，可使用ModelSerializer减少重复代码：

class ContractSerializer(serializers.ModelSerializer):
    class Meta:
        model = Contract
        fields = ['guid', 'number']
        extra_kwargs = {
            'guid': {'label': 'contract_guid'},
            'number': {'label': 'contract_number'}
        }

class ProjectSerializer(serializers.ModelSerializer):
    contracts = ContractSerializer(many=True, source='contract_set')
    
    class Meta:
        model = Project
        fields = ['guid', 'name', 'contracts']
        extra_kwargs = {
            'guid': {'label': 'project_guid'},
            'name': {'label': 'project_name'}
        }

这样就能得到按项目分组的目标输出格式。

内容的提问来源于stack exchange，提问作者serg