求助：使用JOLT将JSON嵌套对象转为数组适配NiFi处理

我有一个包含嵌套对象的JSON负载，需要将其中的嵌套对象修改为数组格式，以便NiFi处理数组元素并写入SQL。

输入JSON #

[
  {
    "architecture": "x64",
    "hostname": "LAPTOP-B442R49A",
    "memory": "32",
    "network0": {
      "asset_id": 1,
      "defaultgateway": "192.168.1.99",
      "dnsserver": "192.168.1.99",
      "ipaddress": "192.168.1.239"
    },
    "network1": {
      "asset_id": 1,
      "defaultgateway": "",
      "dnsserver": "",
      "ipaddress": "192.168.56.1"
    },
    "volume0": {
      "asset_id": 1,
      "caption": "C:\\",
      "freespace": "89"
    },
    "volume1": {
      "asset_id": 1,
      "caption": "D:\\",
      "freespace": "0"
    },
    "asset_id": 1
  }
]

期望输出JSON #

[
  {
    "architecture": "x64",
    "hostname": "LAPTOP-B442R49A",
    "memory": "32",
    "networks": [
      {
        "name": "network0",
        "asset_id": 1,
        "defaultgateway": "192.168.1.99",
        "dnsserver": "192.168.1.99",
        "ipaddress": "192.168.1.239"
      },
      {
        "name": "network1",
        "asset_id": 1,
        "defaultgateway": "",
        "dnsserver": "",
        "ipaddress": "192.168.56.1"
      }
    ],
    "volumes": [
      {
        "name": "volume0",
        "asset_id": 1,
        "caption": "C:\\",
        "freespace": "89"
      },
      {
        "name": "volume1",
        "asset_id": 1,
        "caption": "D:\\",
        "freespace": "0"
      }
    ],
    "asset_id": 1
  }
]

尝试过的JOLT规格（未达预期） #

[
  {
    "operation": "shift",
    "spec": {
      "*": {
        "volume*": {
          "caption": "&.description",
          "freespace": "&.capacity_free",
          "name": "&.drive_name",
          "asset_id": "&.asset_id"
        }
      }
    }
  },
  {
    "operation": "shift",
    "spec": {
      "*": "volumes[]"
    }
  }
]

正确的JOLT转换规格 #

[
  {
    "operation": "shift",
    "spec": {
      "*": {
        // 保留顶层非嵌套字段
        "architecture": "[#2].&",
        "hostname": "[#2].&",
        "memory": "[#2].&",
        "asset_id": "[#2].&",
        // 处理network前缀字段，收集到networks数组并添加name属性
        "network*": {
          "@": "[#3].networks[]",
          "$": "[#3].networks[#2].name"
        },
        // 处理volume前缀字段，收集到volumes数组并添加name属性
        "volume*": {
          "@": "[#3].volumes[]",
          "$": "[#3].volumes[#2].name"
        }
      }
    }
  }
]

规格说明 #

1. shift操作：通过路径匹配实现字段重组
   • 顶层的architecture、hostname等字段直接保留到原数组位置（[#2]表示向上两层的数组索引）
   • 对network*和volume*前缀的字段，先将整个对象放入对应数组（@表示当前节点值），再用$获取字段名作为name属性值
   • [#3]表示向上三层定位到顶层数组元素，[#2]表示当前嵌套对象在数组中的索引位置

内容的提问来源于stack exchange，提问作者Marc