在python中将JSON数据从pandas转换为特定的JSON模式/格式

时间:2017-07-13 18:48:42

标签: python json pandas

我在python中有一些JSON数据,如下所示:

>>> print name_frame
... 
               name    name1    name2    name3     name4
Micro inc.      NaN    Jim D  Susan A      NaN       NaN
Vitacore    Billy B      NaN  Sally Q   Mark G       NaN
>>> payload = name_frame.apply(lambda x: [x.dropna()], axis=1).to_json(force_ascii=False)
... 
>>> print payload
... 




   {
"Micro inc.":[{"name1":"Jim D","name2":"Susan A"}],
    "Vitacore":[{"name":"Billy B","name2":"Sally Q","name3":"Mark G"}],

}

我需要它看起来像这样:

finalJSON = { 
    "company":{
        "name": "Micro inc.",
        "founders": {
            "name": "Jim D",
            "name": "Susan A",
            }
    }
    "company":{
        "name": "Vitacore",
        "founders": {
            "name": "Billy B",
            "name": "Sall Q", 
            "name":"Mark G",
        }

有没有人知道我如何完成这些任务的工具,库或一般建议?我需要将每个公司对象作为POST请求发送到API,它需要这种格式。从那里我需要将结果附加到pandas DataFrame。 我认为应该涉及循环JSON数据,向每家公司提交API,获取结果并将其添加到字典中,或者如果可能的话直接添加到Pandas DataFrame

payload= '''a single company from finalJSON'''

#p is a POST Request
p = requests.post((url + '/r'), json=payload, headers=headers)
p.text #<---- gotta go to a Pandas DataFrame 

提前感谢您提供任何帮助或建议

1 个答案:

答案 0 :(得分:1)

finalJSON = []
for company, names in df.iterrows():
    names = ['"{0}"'.format(name) for name in names.dropna().tolist()]
    names_json_str = ('"name": ' if names else '') + ', "name": '.join(names)
    finalJSON.append('"company": {"name": "' + company + '", "founders": {' + names_json_str + '}')
finalJSON = ', '.join(finalJSON)

>>> finalJSON
'"company": {"name": "Micro inc.", "founders": {"name": "Jim D", "name": "Susan A"}, 
 "company": {"name": "Vitacore", "founders": {"name": "Billy B", "name": "Sally Q", "name": "Mark G"}'