Python 层中的数据帧到json

我有一个数据帧：

school_id  city  id    name   batch   roll_no  age 
   xx       ax    1    abc     a1      100     13
   xx       ax    1    dsf     a2      200     45
   xx       ax    2    fas     a1      400     23

我必须将其转换为多层json格式： 这样，它会在顶部给出一条消息作为键。 school_id和city在所有行中都是相同的，因此它们处于顶级，并且 然后按id嵌套。 我的结果应该是这样的，如果嵌套在id列中，效果会更好

{
  "message": "The school Table.",
  "school_id": "xx",
  "city": "ax",
  "ids": [
    {
     "id" : 1,
     "name" : "abc",
     "batch" : "a1",
     "roll_no" : 100,
     "age" : 13
     },
     {
     "id" : 1,
     "name" : "abc",
     "batch" : "a1",
     "roll_no" : 100,
     "age" : 13
      },
    {
     "id" : 1,
     "name" : "dsf",
     "batch" : "a2",
     "roll_no" : 200,
     "age" : 45
      },
    {
     "id" : 2,
     "name" : "fas",
     "batch" : "a1",
     "roll_no" : 400,
     "age" : 23
      }
       ]
}

---

首先添加新列

message

，然后使用自定义函数为

dict

by添加，最后通过以下方式转换为

json

：

---

基本上，您希望使用groupby：

import json

outputs = []
ids = ["school_id","city"]

for ind,df in df.groupby(ids):
    output = dict(zip(ids,ind),
                  message="School Table",
                  ids=df.drop(ids,axis=1).to_dict('r'))
    outputs.append(output)

print(json.dumps(outputs[0],indent=2))

产出：

{
  "city": "ax",
  "message": "School Table",
  "ids": [
    {
      "roll_no": 100,
      "id": 1,
      "age": 13,
      "batch": "a1",
      "name": "abc"
    },
    {
      "roll_no": 200,
      "id": 1,
      "age": 45,
      "batch": "a2",
      "name": "dsf"
    },
    {
      "roll_no": 400,
      "id": 2,
      "age": 23,
      "batch": "a1",
      "name": "fas"
    }
  ],
  "school_id": "xx"
}

---

“消息”：“学校表。”

手动添加到json中？@jezrael是的……该死的，你太快了@安东：他确实是他不断改进解决方案：）。最多只能投1票though@AntonvBR-谢谢。@jezrael我正在尝试您的数据集代码，您以前的解决方案（执行groupby，然后分配新的列d['message']工作正常，但现在您使用df.assign，我无法在结果json中看到“message”

{
  "city": "ax",
  "message": "School Table",
  "ids": [
    {
      "roll_no": 100,
      "id": 1,
      "age": 13,
      "batch": "a1",
      "name": "abc"
    },
    {
      "roll_no": 200,
      "id": 1,
      "age": 45,
      "batch": "a2",
      "name": "dsf"
    },
    {
      "roll_no": 400,
      "id": 2,
      "age": 23,
      "batch": "a1",
      "name": "fas"
    }
  ],
  "school_id": "xx"
}