仅当第一个 return 不匹配时才在合并中合并

Question

我有 3 个数据框（df1、df2 和 df3），主要数据框 (df1) 和另外两个数据框，其中包含 1 列，我想将它们带到主数据框。

示例 dfs：

df1 = {'Col_f': ['Georgia', 'Nevada', 'New York', 'Texas', 'Arizona'],
          'Col_g': ['SUV', 'Coupe', 'Wagon', 'Crossover', 'Sedan']}
df1 = pd.DataFrame(df1)

df2 = {'Col_g': ['SUV', '4x4', 'Wagon', 'Truck', 'Sedan'],
 'Objective': ['15%', '13%', '55%', '40.4%', '2.48%']}
df2 = pd.DataFrame(df2)

df3 = {'Col_f': ['Georgia', 'California', 'Pennsylvania', 'Texas', 'Arizona'],
 'Objective': ['15%', '13%', '55%', '40.4%', '2.48%']}
df3 = pd.DataFrame(df3)

df1 df2

和df3

我正在使用以下代码：

df1_new = pd.merge(df1, df2, on = 'Col_g', how = 'left')

Which returns the following df:

df_new = {'Col_f': ['Georgia', 'California', 'Pennsylvania', 'Texas', 'Arizona'],
          'Col_g': ['SUV', 'Coupe', 'Wagon', 'Crossover', 'Sedan'],
         'Objective': ['15%', ' ', '55%', ' ', '2.48%']}

然后对于“Objective”列的两个空字符串，我想继续进行第二次合并（或者 excel vlookup，我猜是相同的），以填写那些空的。

我认为将用于第二次合并的代码：

df1_newer = pd.merge(df_new, df3, on = 'Col_f', how = 'left')

期望的最终输出。

df_newer = {'Col_f': ['Georgia', 'California', 'Pennsylvania', 'Texas', 'Arizona'],
          'Col_g': ['SUV', '4x4', 'Wagon', 'Truck', 'Sedan'],
         'Objective': ['15%', '13%', '55%', '40.4%', '2.48%']}

如有任何建议，我们将不胜感激！

Answer 1

合并三个dataframe后，可以使用mask或np.where条件赋值

df1_new = pd.merge(df1, df2, on='Col_g', how='left')['Objective']
df1_newer = pd.merge(df1, df3, on='Col_f', how='left')['Objective']

print(df1_new)

0      NaN
1      NaN
2      55%
3      NaN
4    2.48%
Name: Objective, dtype: object

print(df1_newer)

0      15%
1      NaN
2      NaN
3    40.4%
4    2.48%
Name: Objective, dtype: object

df1['Objective'] = df1_new.mask(df1_new.isna(), df1_newer)

#or

import numpy as np

df1['Objective'] = np.where(df1_new.isna(), df1_newer, df1_new)

print(df1)

      Col_f      Col_g Objective
0   Georgia    Minivan       15%
1    Nevada      Coupe       NaN
2  New York      Wagon       55%
3     Texas  Crossover     40.4%
4   Arizona      Sedan     2.48%

仅当第一个 return 不匹配时才在合并中合并

Merge inside the merge only if the first doesn't return a match

python

numpy

conditional-statements

dataframe

pandas