Python 将dataframe转换为与groupby嵌套的2级JSON

我有一个小要求，我需要以JSON格式列出给定两周内的所有客户及其金额-

  FortNight       Amount      Customer     Parameter
  Apr-2FN-2018   339632.00    10992     CustomerSales
  Apr-2FN-2018   27282.00     10994     CustomerSales 
  Apr-2FN-2018   26353.00     10995     CustomerSales 
  May-1FN-2018   24797.00     11000     CustomerSales
  May-1FN-2018   21093.00     10990     CustomerSales
  May-2FN-2018   45679.0      10992     CustomerSales
  May-2FN-2018   45459.0      10995     CustomerSales
  May-2FN-2018   35885.0      10994     CustomerSales

预期的JSON-

"CustomerSales" : [                                                                
    {"FortNight" : "Apr-2FN-2018",                                                                                      
         "Details" :[
             {"Customer":  "10992","Amount" : 339632.00},                                                                                                                                
             {"Customer":  "10994","Amount" : 27282.00},
             {"Customer":  "10995","Amount" : 26353.00}
           ]
    },
    {"FortNight" : "May-1FN-2018",                                                                                      
         "Details" :[
             {"Customer":  "11000","Amount" : 24797.00},                                                                                                                                
             {"Customer":  "10990","Amount" : 21093.00}
           ]
    },
    {"FortNight" : "May-2FN-2018",                                                                                      
         "Details" :[
             {"Customer":  "10992","Amount" : 45679.00},                                                                                                                                
             {"Customer":  "10995","Amount" : 45459.00},
             {"Customer":  "10994","Amount" : 35885.00}
           ]
    }
]

我试着从一个角度来理解这个逻辑-

但我仍然停留在另一个层次

---

我也尝试过其他的方法，但没有成功。欢迎任何帮助。提前谢谢

您可以使用

CustomerSales

字典，其中键等于

FortNight

值，而不是在每个字典中使用“FortNight”字符串作为键的字典列表，从而简化此操作。鉴于规定的要求，以下各项应适用。您可以循环遍历数据帧并构建字典。然后，第二组代码将该字典转换为您指定的格式

sales = {}

for idx, row in df.iterrows():
    fn = row["FortNight"]
    fn_sales = sales.get(fn)
    if fn_sales is None:
        sales[fn] = {
                "Details": []
                }
        fn_sales = sales.get(fn)

    fn_sales["Details"].append({
            "Customer": row["Customer"],
            "Amount": row["Amount"]
            }
    )

sales_obj = {"CustomerSales": []}

for fn, details in sales.items():
    sales_obj["CustomerSales"].append({
            "FortNite": fn,
            "Details": details["Details"]
            })

---

什么方法对你不起作用？目前我可以得到如下输出-[{'Customersales'：[{'FortNight'：'Apr-2FN-2018'，'Details'：[{'Customer'：'1100'，'Amount'：198218.02}]}]，{'Customersales'：[{'FortNight'：'Apr-3FN-2018'，'Details'：[{'Customer'：'1100'，'Amount:198218.02}]}]，{'Customersales':[{'FortNight'：'May-1FN-2018'，'Details'：[{'Customer'：'1100'，'Amount'：198218.02}]}]