如何为datafarme创建嵌套和自定义json格式

Question

我希望从现有的数据帧数据框架列中创建子类别(示例表)，我在列级别上所需的更改，而不是像一组列一样的数据中的任何更改，都是和列名称一样的，3个不同的后缀(很少有类似的列名和其他列名)，例如

|payer_id|payer_name|halo_payer_name|delta_payer_name|halo_desc|delta_desc|halo_operations|delta_notes|halo_processed_data|delta_processed_data|extra|insurance_company|

我希望它被分组在这个光环组halo_payer_name|halo_desc|halo_operations|halo_processed_data|中

我希望它被分组在这个增量组delta_payer_name|delta_desc|delta_notes|delta_processed_data|中

和其余的列作为一个组，所以当转换为JSON时，它将出现在下面的布局中

{
    "schema": {
        "fields": [{
                "payer_details": [{
                        "name": "payer_id",
                        "type": "string"
                    },
                    {
                        "name": "payer_name",
                        "type": "string"
                    },
                    {
                        "name": "extra",
                        "type": "string"
                    },
                    {
                        "name": "insurance_company",
                        "type": "string"
                    }
                ]
            },
            {
                "halo": [{
                        "name": "halo_payer_name",
                        "type": "string"
                    },
                    {
                        "name": "halo_desc",
                        "type": "string"
                    },
                    {
                        "name": "halo_operstions",
                        "type": "string"
                    },
                    {
                        "name": "halo_processed_data",
                        "type": "string"
                    }
                ]
            }, {
                "delta": [{
                        "name": "delta_payer_name",
                        "type": "string"
                    },
                    {
                        "name": "delta_desc",
                        "type": "string"
                    },
                    {
                        "name": "delta_notes",
                        "type": "string"
                    },
                    {
                        "name": "delta_processed_data",
                        "type": "string"
                    }
                ]
            }
        ],
        "pandas_version": "1.4.0"
    },
    "masterdata": [{
        "payer_details": [{
            "payer_id": "",
            "payer_name": "",
            "extra": "",
            "insurance_company": ""
        }],
        "halo": [{
            "halo_payer_name": "",
            "halo_desc": "",
            "halo_operations": "",
            "halo_processed_data": "",
                    }],
        "delta":[{
            "delta_payer_name": "",
            "delta_desc": "",
            "delta_notes": "",
            "delta_processed_data": "",
                    }]
    }]
}

对于这种情况，我找不到解决方案，因为它是基于列的分组，而不是基于数据的分组。

Answer 1

所以今天看到这篇文章，它帮助我解决了我的问题(从数据框架中添加数据，并使用它创建循环数据并将其插入到dict中，然后将其转换为JSON文件)，对我有帮助的参考是link，所以这个问题的解决方案如下

schema={
    "schema": {
        "fields": [{
                "payer_details": [{
                        "name": "payer_id",
                        "type": "string"
                    },
                    {
                        "name": "payer_name",
                        "type": "string"
                    },
                    {
                        "name": "extra",
                        "type": "string"
                    },
                    {
                        "name": "insurance_company",
                        "type": "string"
                    }
                ]
            },
            {
                "halo": [{
                        "name": "halo_payer_name",
                        "type": "string"
                    },
                    {
                        "name": "halo_desc",
                        "type": "string"
                    },
                    {
                        "name": "halo_operstions",
                        "type": "string"
                    },
                    {
                        "name": "halo_processed_data",
                        "type": "string"
                    }
                ]
            }, {
                "delta": [{
                        "name": "delta_payer_name",
                        "type": "string"
                    },
                    {
                        "name": "delta_desc",
                        "type": "string"
                    },
                    {
                        "name": "delta_notes",
                        "type": "string"
                    },
                    {
                        "name": "delta_processed_data",
                        "type": "string"
                    }
                ]
            }
        ],
        "pandas_version": "1.4.0"
    },
    "masterdata": []
}

按我所希望的方式导出上述架构。

payer_list=[]
for i in df.index:
  case={
        "payer_details": [{
            "payer_id": "{}".format(df['payer_id'][i]),
            "payer_name": "{}".format(df['payer_name'][i]),
            "extra": "{}".format(df['extra'][i]),
            "insurance_company": "{}".format(df['insurance_company'][i])
        }],
        "halo": [{
            "halo_payer_name": "{}".format(df['halo_payer_name'][i]),
            "halo_desc": "{}".format(df['halo_desc'][i]),
            "halo_operations": "{}".format(df['halo_operations'][i]),
            "halo_processed_data": "{}".format(df['halo_processed_data'][i]),
                    }],
        "delta":[{
            "delta_payer_name": "{}".format(df['delta_payer_name'][i]),
            "delta_desc": "{}".format(df['delta_desc'][i]),
            "delta_notes": "{}".format(df['delta_notes'][i]),
            "delta_processed_data": "{}".format(df['delta_processed_data'][i]),
                    }]
    }
  payer_list.append(case)
schema["masterdata"] = payer_list

创建和空列表并运行循环，并包含在空列表中，并连接或链接到架构。