匹配两个不同DataFrame中的字符串值，并在Pandas中创建一个带匹配指示符的新列

Question

我有两个数据帧（df1，df2），我想 在df1中创建一个新列，指示是否 每个代码列之间都有一个匹配项 数据帧。 df2中的代码列由 用逗号分隔的字符串。

DF1

Date        Code
2016-01-01  LANH08
2016-01-01  LAOH07
2016-01-01  LAPH09
2016-01-01  LAQH06
2016-01-01  LARH03

DF2

Date        Code
2016-01-01  LANH08, LAOH07, LXA0EW, LAGRL1
2016-01-01  LAUH02, LAVH00, LAVH01, LAYH00
2016-01-01  LANH08
2016-01-01  AAH00, ABH00, XAH03
2016-01-01  ARH04, BA0BW, BMH01, DPH00

我的目标

DF1

Date        Code    Match
2016-01-01  LANH08  Y
2016-01-01  LAOH07  Y
2016-01-01  LAPH09  N
2016-01-01  LAQH06  N
2016-01-01  LARH03  N

Answer 1

#Split df2['Code'] into an array
df2.Code = df2.Code.str.split(', ')

#Recreate df2 reshaped
df2 = pd.concat([pd.DataFrame(dict(list(zip(df2.columns,df2.ix[i]))),\
      index=range(len(list(zip(df2.columns,df2.ix[i]))[1]))) for i in range(len(df2.index))])


#default df2['match'] to 'Y'
df2['Match'] = 'Y'

#Create new dataframe by left merging df1 with df2
df3 = df1.merge(df2, left_on = ['Date','Code'], right_on = ['Date','Code'], how = 'left')

#Fill NaN values in Match column with 'N' (because they weren't in df2)
df3['Match'] = df3['Match'].fillna('N')

Answer 2

Split pandas dataframe string entry to separate rows

最终解决方案：

data1 = {'Date':['2016-01-01',
                 '2016-01-01',
                 '2016-01-01',
                 '2016-01-01',
                 '2016-01-01'],
         'Code':['LANH08',
                 'LAOH07',
                 'LAPH09',
                 'LAQH06',
                 'LARH03']}
df1 = DataFrame(data1)

data2 = {'Date':['2016-01-01',
                 '2016-01-01',
                 '2016-01-01',
                 '2016-01-01',
                 '2016-01-01'],
         'Code':['LANH08, LAOH07, LXA0EW, LAGRL1',
                 'LAUH02, LAVH00, LAVH01, LAYH00',
                 'LANH08',
                 'AAH00, ABH00, XAH03',
                 'LAUH02, LAVH00']}
df2 = DataFrame(data2)

df2 = DataFrame(df2.Code.str.split(', ').tolist(), index=df2.Date).stack().drop_duplicates()
df2 = df2.reset_index()[[0, 'Date']] # Code variable is currently labeled 0
df2.columns = ['Code', 'Date'] # Renaming Code

# default df2['match'] to 'Y'
df2['Match'] = 'Y'

# Create new dataframe by left merging df1 with df2
df3 = df1.merge(df2, left_on = ['Code', 'Date'], right_on = ['Code', 'Date'], how = 'left')

# Fill NaN values in Match column with 'N' (because they weren't in df2)
df3['Match'] = df3['Match'].fillna('N')

df3

    Code    Date        Match
0   LANH08  2016-01-01  Y
1   LAOH07  2016-01-01  Y
2   LAPH09  2016-01-01  N
3   LAQH06  2016-01-01  N
4   LARH03  2016-01-01  N