如何将几列熊猫数据框转换为另一条记录中的 JSON 记录？

Posted 2023-03-12

【中文标题】如何将几列熊猫数据框转换为另一条记录中的 JSON 记录？

【英文标题】：How to convert few columns of pandas dataframe into JSON record within another record?

【发布时间】：2016-03-19 17:10:27

【问题描述】：

所以我正在尝试将具有 12 列以上的 pandas 数据框转换为相应的 JSON 记录。我能得到它。但是，我希望框架的几列成为新列的子记录。这怎么能实现。？

                    "ADRNR": 2692629, 
                    "AlertID": "", 
                    "AlertTimestamp": "14-12-2015 14:44:14", 
                    "BANKL": null, 
                    "BANKN": null, 
                    "BANKS": "nan", 
                    "BEGRU": "NPIV", 
                    "BUKRS": "2646", 
                    "C_Block": "No", 
                    "KOINH": null, 
                    "LAND1": "US", 
                    "LOEVM_x": null, 
                    "LOEVM_y": null, 
                    "MasterDataID": "10099", 
                    "MasterDataType": "Vendor", 
                    "NAME1": "LEGAL", 
                    "NODEL_x": null, 
                    "NODEL_y": null, 
                    "ORT01": null, 
                    "OtherData": null, 
                    "PSTLZ": null, 
                    "RuleID": "Rule3", 
                    "RuleName": "Vendor and Bank Country is Different", 
                    "STCD1": null, 
                    "STCD2": null, 
                    "STCEG": null, 
                    "STRAS": null, 
                    "TELF1": null
               

上面的 JSON 是我得到的。但我想要以下结构。请指导我。

 

                "RuleID": "Rule3", 
                "RuleName": "Vendor and Bank Country is Different", 
                "AlertID": "", 
                "AlertTimestamp": "14-12-2015 14:44:14", 
                "MasterDataID": "10099", 
                "MasterDataType": "Vendor", 
                "OtherData": 
                    "BANKL": null, 
                    "BANKN": null, 
                    "BANKS": "nan", 
                    "BEGRU": "NPIV", 
                    "BUKRS": "2646", 
                    "C_Block": "No", 
                    "KOINH": null, 
                    "LAND1": "US", 
                    "LOEVM_x": null, 
                    "LOEVM_y": null, 
                    "NAME1": "LEGAL", 
                    "NODEL_x": null, 
                    "NODEL_y": null, 
                    "ORT01": null,
                    "PSTLZ": null, 
                    "ADRNR": 2692629,
                    "STCD1": null, 
                    "STCD2": null, 
                    "STCEG": null, 
                    "STRAS": null, 
                    "TELF1": null
                
           

编辑：Flg 是我的代码

Final_Table['AlertID'] = ''
Final_Table['AlertTimestamp'] = datetime.now().strftime('%d-%m-%Y %H:%M:%S')
Final_Table['MasterDataType'] = 'Vendor'
Final_Table['RuleID'] = 'Rule3'
Final_Table['RuleName'] = 'Vendor and Bank Country is Different'
Final_Table = Final_Table.rename(columns='LIFNR': 'MasterDataID')


Result = Final_Table[Final_Table['BANKS'] != Final_Table['LAND1']]

Result['OtherData'] = np.NaN


final_result = 'alerts': json.loads(Result.to_json(orient = 'records',force_ascii = False).encode('utf8'))
result = 'results': final_result

with open('output_Rule3.json', 'w') as outfile:
    json.dump(result, outfile, indent = 5, sort_keys = True)
log.info("Rule3 : Execution Successful")

【问题讨论】：

你能展示你的努力吗

@EdChum 添加了代码！

你想要的和你得到的有什么区别？它们看起来很相似。

@JohnZwinck 我想要几列进入“OtherData”。如果您比较两个 JSON 中的 OtherData，您会注意到差异。

【参考方案1】：

我将修改json 输出，而不是在pandas 内。

从您的 json 数据开始

print(json.dumps(data, indent=4, sort_keys=True))

[
    
        "ADRNR": 2692629,
        "AlertID": "",
        "AlertTimestamp": "14-12-2015 14:44:14",
        "BANKL": null,
        "BANKN": null,
        "BANKS": "nan",
        "BEGRU": "NPIV",
        "BUKRS": "2646",
        "C_Block": "No",
        "KOINH": null,
        "LAND1": "US",
        "LOEVM_x": null,
        "LOEVM_y": null,
        "MasterDataID": "10099",
        "MasterDataType": "Vendor",
        "NAME1": "LEGAL",
        "NODEL_x": null,
        "NODEL_y": null,
        "ORT01": null,
        "OtherData": null,
        "PSTLZ": null,
        "RuleID": "Rule3",
        "RuleName": "Vendor and Bank Country is Different",
        "STCD1": null,
        "STCD2": null,
        "STCEG": null,
        "STRAS": null,
        "TELF1": null
    
]

以及您感兴趣的一些columns 的列表：

other_data = ["BANKL", "BANKN", "BANKS", "BEGRU", "BUKRS", "C_Block"]

你可以直接修改json：

for i, record in enumerate(data):
    data[i]['OtherData'] = 
    for key in list(record.keys()):
        if key in other_data:
            data[i]['OtherData'][key] = record[key]
            del data[i][key]

print(json.dumps(data, indent=4, sort_keys=True))

[
    
        "ADRNR": 2692629,
        "AlertID": "",
        "AlertTimestamp": "14-12-2015 14:44:14",
        "KOINH": null,
        "LAND1": "US",
        "LOEVM_x": null,
        "LOEVM_y": null,
        "MasterDataID": "10099",
        "MasterDataType": "Vendor",
        "NAME1": "LEGAL",
        "NODEL_x": null,
        "NODEL_y": null,
        "ORT01": null,
        "OtherData": 
            "BANKL": null,
            "BANKN": null,
            "BANKS": "nan",
            "BEGRU": "NPIV",
            "BUKRS": "2646",
            "C_Block": "No"
        ,
        "PSTLZ": null,
        "RuleID": "Rule3",
        "RuleName": "Vendor and Bank Country is Different",
        "STCD1": null,
        "STCD2": null,
        "STCEG": null,
        "STRAS": null,
        "TELF1": null
    
]

【讨论】：

感谢您的帮助。它确实奏效了。在将其转换为 JSON 之前，我一直在努力对其进行格式化。