Question

我有一个数据框（'sp500news'），如下所示：

date_publish  \
79944   2007-01-29 19:08:35   
181781  2007-12-14 19:39:06   
213175  2008-01-22 11:17:19   
93554   2008-01-22 18:52:56   
  ...

title  
 79944   Microsoft Vista corporate sales go very well                                            
 181781  Williams No Anglican consensus on Episcopal Church                                      
 213175  CSX quarterly profit rises                                                              
 93554   Citigroup says 30 bln capital helps exceed target                                       
    ...

我正在尝试通过df的“符号”列（“组成部分”）中的相应代码更新每个公司名称，如下所示：

Symbol  Name    Sector
0   MMM 3M  Industrials
1   AOS A.O. Smith  Industrials
2   ABT Abbott  Health Care
3   ABBV    AbbVie  Health Care
...
116  C      Citigroup    Financials       
...

我已经尝试过：

for item in sp500news['title']:
    for word in item:
        if word in constituents['Name']:
            indx = constituents['Name'].index(word)
            str.replace(word, constituents['Symbol'][indx])

Answer 1

尝试一下：

以下是代表您的数据的虚拟数据框

df1 = pd.DataFrame({'Symbol': ['MV', 'AOS','ABT'],
                  'Name': ['Microsoft Vista', 'A.0.', 'Abbot']})
df1
  Symbol    Name
0   MV  Microsoft Vista
1   AOS A.0.
2   ABT Abbot
df2 = pd.DataFrame({'title': [79944, 181781, 213175],
                   'comment': ['Microsoft Vista corporate sales go very well',
                              'Abbot consensus on Episcopal Church',
                              'A.O. says 30 bln captial helps exceed target']})

    title   comment
0   79944   Microsoft Vista corporate sales go very well
1   181781  Abbot consensus on Episcopal Church
2   213175  A.O. says 30 bln captial helps exceed target

制作一个值字典，将名称映射到它们各自的符号

rep = dict(zip(df1.Name,df1.Symbol))
rep

{'Microsoft Vista': 'MV', 'A.0.': 'AOS', 'Abbot': 'ABT'}

使用Series.replace方法替换它们

df2['comment'] = df2['comment'].replace(rep, regex = True)
df2
   title    comment
0   79944   MV corporate sales go very well
1   181781  ABT consensus on Episcopal Church
2   213175  A.O. says 30 bln captial helps exceed target

Answer 2

尝试以下代码

df = pd.DataFrame({'title': ['Citigroup says 30 bln capital helps exceed target',
                             'Williams No Anglican consensus on Episcopal Church',
                             'Microsoft Vista corporate sales go very well']})

constituents = pd.DataFrame({'symbol': ['MMM', 'C', 'MCR', 'WLM'],
                             'name': ['3M', 'Citigroup', 'Microsoft', 'Williams']})

for name, symbol in zip(constituents['name'], constituents['symbol']):
    df['title'] = df['title'].str.replace(name, symbol)

输出

                                           title
0      C says 30 bln capital helps exceed target
1  WLM No Anglican consensus on Episcopal Church
2         MCR Vista corporate sales go very well

我基本上只是复制了sp500news['title]的几行并组成了constituents['Name']的一部分，只是为了演示转换。本质上，我正在从pd.Series访问列title的{{1}}对象的sp500news对象的字符串方法对象，因此当找到匹配的公司名称时，可以对其应用replace

如何在字符串的Pandas DataFrame列中更新子字符串

2 个答案: