解析JSON行fi

Question

我需要找到一种将json文件中的数据解析为csv或xlsx的方法。 但是，我在线使用的每个JSON验证器都会给我一个错误，说JSON文件无效。在

JSON文件示例如下：

{"id": "someID1.docx",
 "language": {"detected": "cs"},
 "title": "Name - Title - FileName",
 "text": "Long string of text",
 "entities": [
 {"standardForm": "Svářečský průkaz", "type": "car"},
 {"standardForm": "email1@gmail.com", "type": "email"},
 {"standardForm": "english", "type": "languages"},
 {"standardForm": "Práce na PC", "type": "abilities"},
 {"standardForm": "MS Office", "type": "abilities"},
 {"standardForm": "Automechanik", "type": "education"},
 {"standardForm": "Střední průmyslová škola", "type": "education"},
 {"standardForm": "Angličtina-Němčina", "type": "languages"},
 {"standardForm": "mechanic", "type": "position"},
 {"standardForm": "Praha", "type": "region"},
 {"standardForm": "B2 - středně pokročilý", "type": "en_level"},
 {"standardForm": "Skupina B", "type": "drivinglicense"}
 ]}
{"id": "someID2.pdf",
 "language": {"detected": "cs"},
 "title": "Name - Title - FileName2",
 "text": "Long string of text2",
 "entities": [
 {"standardForm": "german", "type": "languages"},
 {"standardForm": "high school", "type": "education"},
 {"standardForm": "Angličtina-Němčina", "type": "languages"},
 {"standardForm": "driver", "type": "position"},
 {"standardForm": "english", "type": "languages"},
 {"standardForm": "university", "type": "education"},
 {"standardForm": "email2@seznam.cz", "type": "email"},
 {"standardForm": "Středočeský", "type": "region"},
 {"standardForm": "Střední", "type": "edulevel"},
 {"standardForm": "manager", "type": "lastposition"},
 {"standardForm": "? – nerozpoznáno", "type": "de_level"},
 {"standardForm": "? – nerozpoznáno", "type": "en_level"},
 {"standardForm": "Skupina C", "type": "drivinglicense"}
 ]}
 ...

我可以用Python加载这个JSON：

^{pr2}$

但我无法将其转换成csv。我需要能够存储所有与所有id相关的实体，最好是csv。有什么办法吗？我需要JSON不同吗？在

谢谢

编辑： 我需要上面例子的csv输出如下：

ID;title;languages;education
someID1.docx;Name-Title-FileName;english,Angličtina-Němčina;Automechanik;Střední Prům. škola
seomeID2.pdf;Name-Title-FileName2; german,Angličtina-Němčina,english;high school, university

Answer 1

用miller（https://github.com/johnkerl/miller/releases/tag/5.4.0），简单地

mlr j2c unsparsify then cut -x -r -f "entit" input.json >output.csv

你有这个CSV

id,language:detected,title,text
someID1.docx,cs,Name - Title - FileName,Long string of text
someID2.pdf,cs,Name - Title - FileName2,Long string of text2

关于选项的一些注意事项：

• j2c将json转换为csv
• unsparsify在所有输入记录上打印字段名并集的记录
• cut -x -r -f从JSON中删除entities对象

Answer 2

你已经可以雇佣它的熊猫了

df = pd.DataFrame(jsonfile)
df['languages'] = df.apply(lambda x: [item['standardForm'] 
                                      for item in x.entities 
                                      if item['type'] == 'languages'], 
                           axis=1)
df['education'] = df.apply(lambda x: [item['standardForm'] 
                                      for item in x.entities 
                                      if item['type'] == 'education'],
                           axis=1)


df.to_csv(<filename>, columns=['id', 'title', 'languages', 'education'])