将列中的值提取到新列答案

【问题标题】：Extract value in column to new columns将列中的值提取到新列
【发布时间】：2020-11-25 05:20:11
【问题描述】：

我想将列内部提取到几列。这是导入数据框后的原始数据。

data = {'ID':  ['A0001', 'A0002', 'A0003', 'A0004', 'A0005'],
        'Name': ['John', 'Micheal', 'Angle', 'Jim', 'Rome'],
        'Details': ['Type:\nHouse\nVector:\nTriangle\n\nMission:\nCompleted,lv5\n\nNote user:\n#', 'Type:\n#\nVector:\n\n\nMission:\nFailed\nNote user:\n#', 'Type:\nCar\nVector:\nSquare\nMission:\nCompleted\nNote user:\n', 'Type:\n#\nVector:\n#\nMission:\nCompleted without award\n\nNote user:\nNo end', 'Type:\n#\nVector:\n#\nMission:\n\n\nNote user:\nThere are many mistake.\nI cant choose.\nI cant buy.']
        }

df = pd.DataFrame (data, columns=['ID', 'Name', 'Details'])
df

ID      Name    Details
A0001   John    Type:\nHouse\nVector:\nTriangle\n\nMission:\nCompleted,lv5\n\nNote user:\n#
A0002   Micheal Type:\n#\nVector:\n\n\nMission:\nFailed\nNote user:\n#
A0003   Angle   Type:\nCar\nVector:\nSquare\nMission:\nCompleted\nNote user:\n
A0004   Jim     Type:\n#\nVector:\n#\nMission:\nCompleted without award\n\nNote user:\nNo end
A0005   Rome    Type:\n#\nVector:\n#\nMission:\n\n\nNote user:\nThere are many mistake.\nI cant choose.\nI cant buy.

我想提取详细信息列中的值。但我不知道该怎么做。

我的预期数据看起来像

data = {'ID':  ['A0001', 'A0002', 'A0003', 'A0004', 'A0005'],
        'Name': ['John', 'Micheal', 'Angle', 'Jim', 'Rome'],
        'Type': ['House', '#', 'Car', '#', '#'],
        'Vector': ['Triangle', '', 'Square', '#', '#'],
        'Mission': ['Completed,lv5', 'Failed', 'Completed', 'Completed without award', ''],
        'Note user': ['#', '#', '', 'No end', 'There are many mistake.I cant choose.I cant buy.']
        }

df = pd.DataFrame (data, columns=['ID', 'Name', 'Type', 'Vector', 'Mission', 'Note user'])
df



ID      Name    Type    Vector      Mission                      Note
A0001   John    House   Triangle    Completed,lv5                #
A0002   Micheal #                   Failed                       #
A0003   Angle   Car     Square      Completed   
A0004   Jim     #       #           Completed without award      No end
A0005   Rome    #       #                                        There are many mistake.I cant choose.I cant buy.

【问题讨论】：

看起来您想按\n 拆分列并将所有值存储到不同的列中
类似但有一些值\n\n\n标题也是带冒号的单词（例如，Type: Vector: Mission: Note:)。

标签： python pandas dataframe data-extraction

【解决方案1】：

这是我尝试过的：Details 中的第一个值是：

'Type:\nHouse\nVector:\nTriangle\n\nMission:\nCompleted,lv5\n\nNote user:\n#'

我编写了这个函数来提取细节到dict。我对数组的索引进行了硬编码，但如果您愿意，可以选择不这样做：

def extract_details(text):
    array = text.replace("\n\n", "\n").split("\n")
    return {
        array[0].replace(":", ""): array[1],
        array[2].replace(":", ""): array[3],
        array[4].replace(":", ""): array[5],
        array[6].replace(":", ""): array[7]
    }

将函数应用于整个列：

df['Details'].apply(extract_details)

将此新列连接到原始列：

pd.concat([
    df,
    pd.DataFrame(df['Details'].apply(extract_details).apply(pd.Series))
], axis=1)

【讨论】：

【解决方案2】：

您可以使用regex 来获取答案。附上文档链接。

首先，我将所有\n 替换为''。这样一来，所有换行符都会从 Details 列中删除。

然后我抓取两个关键字之间的所有文本。对于类型，数据介于'Type:' 和'Vector:' 之间。 Vector 和 Mission 也是如此。注意，我正在获取'Note user:' 之后的所有数据。现在您已经从Details 列中提取了数据，您可以删除该列。

import pandas as pd
data = {'ID':  ['A0001', 'A0002', 'A0003', 'A0004', 'A0005'],
        'Name': ['John', 'Micheal', 'Angle', 'Jim', 'Rome'],
        'Details': ['Type:\nHouse\nVector:\nTriangle\n\nMission:\nCompleted,lv5\n\nNote user:\n#', 'Type:\n#\nVector:\n\n\nMission:\nFailed\nNote user:\n#', 'Type:\nCar\nVector:\nSquare\nMission:\nCompleted\nNote user:\n', 'Type:\n#\nVector:\n#\nMission:\nCompleted without award\n\nNote user:\nNo end', 'Type:\n#\nVector:\n#\nMission:\n\n\nNote user:\nThere are many mistake.\nI cant choose.\nI cant buy.']
        }

df = pd.DataFrame (data, columns=['ID', 'Name', 'Details'])

df['Details'] = df.Details.str.replace('\n','', regex=True)
df['Type']    = df.Details.str.extract('Type\:(.*)Vector')
df['Vector']  = df.Details.str.extract('Vector\:(.*)Mission')
df['Mission'] = df.Details.str.extract('Mission\:(.*)Note')
df['Note']    = df.Details.str.extract('Note user\:(.*)')

print (df[['ID','Name','Type','Vector']])
print (df[['Mission','Note']])

这个输出将是：

      ID     Name   Type    Vector
0  A0001     John  House  Triangle
1  A0002  Micheal      #          
2  A0003    Angle    Car    Square
3  A0004      Jim      #         #
4  A0005     Rome      #         #

                   Mission                                              Note
0            Completed,lv5                                                 #
1                   Failed                                                 #
2                Completed                                                  
3  Completed without award                                            No end
4                           There are many mistake.I cant choose.I cant buy.

【讨论】：