열이있는 텍스트 파일을 CSV로 변환

1. 행이 msg sep 행을 포함하는 것을 찾습니다. 예 : '-----', '======'
2. 그런 다음 np.where(cond, 1, 0).cumsum()모든 개별 메시지에 태그를 지정 하는 데 사용하십시오 .
3. '-----'또는 '======'없이 행 필터링
4. groupby 태그를 지정하고 sep '\ n'으로 조인 한 다음 str.split을 사용하여 열을 확장합니다.

# read the file with only one col
df = pd.read_csv(file, sep='\n', header=None)

# located the row contains ------ or ======
cond = df[0].str.contains('-----|======')
df['tag'] = np.where(cond, 1, 0).cumsum()

# filter the line contains msg
cond2 = df['tag'] >=2
dfn = df[(~cond & cond2)].copy()

# output
df_output = (dfn.groupby('tag')[0]
            .apply('\n'.join)
            .str.split('\n', n=2, expand=True))
df_output.columns = ['name', 'date', 'Description']

산출:

              name                            date  \
tag                                                  
2.0        1: Aki        2018/12/05 (Wed) 17:33:17   
3.0         2: Aki  2018/12/06 (Thursday) 17:14:30   
4.0  3: kano, etc.    2018/12/07 (Friday) 11:44:45   

                                           Description  
tag                                                     
2.0                  An approval notice has been sent.  
3.0  I was notified by Mr. Id, the agent of the oth...  
4.0                                  Please call rito.  

df :

                                                    0  tag
0   ==============================================...    1
1                                   Title: Whole case    1
2                                     Location: oyuri    1
3                                          From: Aki     1
4                  Date: 2018/11/30 (Friday) 11:55:29    1
5   ==============================================...    2
6                                             1: Aki     2
7                           2018/12/05 (Wed) 17:33:17    2
8                   An approval notice has been sent.    2
9   ----------------------------------------------...    3
10                                             2: Aki    3
11                     2018/12/06 (Thursday) 17:14:30    3
12  I was notified by Mr. Id, the agent of the oth...    3
13  ----------------------------------------------...    4
14                                      3: kano, etc.    4
15                       2018/12/07 (Friday) 11:44:45    4
16                                  Please call rito.    4
17  ----------------------------------------------...    5

계속해서 이름을 처리 할 수 ​​있습니다.

obj = df_output['name'].str.strip().str.split(':\s*')
df_output['name'] = obj.str[-1]
df_output['idx'] = obj.str[0]
df_output = df_output.set_index('idx')

           name                            date  \
idx                                               
1           Aki       2018/12/05 (Wed) 17:33:17   
2           Aki  2018/12/06 (Thursday) 17:14:30   
3    kano, etc.    2018/12/07 (Friday) 11:44:45   

                                           Description  
idx                                                     
1                    An approval notice has been sent.  
2    I was notified by Mr. Id, the agent of the oth...  
3                                    Please call rito.

더 많은 헤더 열 추가 :

cond = (df['tag'] == 1) & (df[0].str.contains(':'))
header_dict = dict(df.loc[cond, 0].str.split(': ', n=1).values)

    # {'Title': 'Whole case',
    #  'Location': 'oyuri',
    #  'From': 'Aki ',
    #  'Date': '2018/11/30 (Friday) 11:55:29'}

for k,v in header_dict.items():
    df_output[k] = v