2 嵌套字典 Json 格式使用 Json_norminalize 使用 python

Question

您好，我正在尝试检索此数据，其中有 2 个嵌套字典：

{
"metadata": {
    "stations": [
        {
            "id": "S108",
            "device_id": "S108",
            "name": "Kuala Lumpur",
            "location": {
                "latitude": 3.1390,
                "longitude": 101.6869
            }
        },
        {
            "id": "S118",
            "device_id": "S118",
            "name": "Bukit Bintang",
            "location": {
                "latitude":3.1468,
                "longitude": 101.7113
            }
        }
    ],
    "reading_type": "DBT 1M F",
    "reading_unit": "deg C"
},
"items": [
    {
        "timestamp": "2021-06-20T15:05:00+08:00",
        "readings": [
            {
                "station_id": "S108",
                "value": 32.6
            },
            {
                "station_id": "S118",
                "value": 30.3
            }
        ]
    }
]

我想得到这样的结果： Result

我尝试了几种方法：

data = airtemp.json()
df = pd.json_normalize(data,record_path=['metadata', 'stations'])
df


data = airtemp.json()
df1 = pd.json_normalize(data,record_path=['items','readings'])
df1

有没有一种方法可以使用 json_norminalize 形成一个包含 station_id、名称、纬度、经度、时间戳和值的表，而不会分成 2 个表？

非常感谢！

Answer 1

你可以merge你创建的两个数据框：

import pandas as pd


data = {
  "metadata": {
    "stations": [
      {
        "id": "S108",
        "device_id": "S108",
        "name": "Kuala Lumpur",
        "location": {
          "latitude": 3.1390,
          "longitude": 101.6869
        }
      },
      {
        "id": "S118",
        "device_id": "S118",
        "name": "Bukit Bintang",
        "location": {
          "latitude": 3.1468,
          "longitude": 101.7113
        }
      }
    ],
    "reading_type": "DBT 1M F",
    "reading_unit": "deg C"
  },
  "items": [
    {
      "timestamp": "2021-06-20T15:05:00+08:00",
      "readings": [
        {
          "station_id": "S108",
          "value": 32.6
        },
        {
          "station_id": "S118",
          "value": 30.3
        }
      ]
    }
  ]
}

stations = pd.json_normalize(data, record_path=['metadata', 'stations'])
readings = pd.json_normalize(data, record_path=['items', 'readings'])
result = stations.merge(readings, left_on='id', right_on='station_id')
with pd.option_context('display.max_rows', None, 'display.max_columns', None):
    print(result)

输出：

   id      device_id     name    location.latitude  location.longitude  \
0  S108      S108   Kuala Lumpur             3.1390            101.6869   
1  S118      S118  Bukit Bintang             3.1468            101.7113   

     station_id  value  
0       S108   32.6  
1       S118   30.3  

您提供的数据中只有一个timestamp，因此您必须单独获取它。