Question

我想写一个函数，当df1和df2的列名相互匹配时，它会更新df1的值。

例如： df1：

    Name | Graduated | Employed | Married
    AAA       1           2         3
    BBB       0           1         2 
    CCC       1           0         1

df2：

    Answer_Code | Graduated | Employed | Married
       0                No         No        No
       1                Yes       Intern    Engaged
       2                N/A        PT        Yes
       3                N/A        FT      Divorced

最终结果： DF3：

     Name | Graduated |   Employed   |  Married
     AAA       Yes          PT         Divorced
     BBB       No           Intern     Yes 
     CCC       Yes          No         NO

我想编写类似这样的代码：

     IF d1.columns = d2.columns THEN 

     df1.column.update(df1.column.map(df2.set_index('Answer_Code').column))

Answer 1

您可以使用map。

示例：

df1.Graduated.map(df2.Graduated)

产量

0    Yes
1     No
2    Yes

因此，只需对每一列执行此操作，如下所示

for col in df1.columns:
    if col in df2.columns:
        df1[col] = df1[col].map(df2[col])

请务必先将索引设置为答案代码，如有必要，请设置df2 = df2.set_index("Answer_Code")。

Answer 2

一种方法是使用pd.DataFrame.lookup：

df1 = pd.DataFrame({'Name': ['AAA', 'BBB', 'CCC'],
                    'Graduated': [1, 0, 1],
                    'Employed': [2, 1, 0],
                    'Married': [3, 2, 1]})

df2 = pd.DataFrame({'Answer_Code': [0, 1, 2, 3],
                    'Graduated': ['No', 'Yes', np.nan, np.nan],
                    'Employed': ['No', 'Intern', 'PT', 'FT'],
                    'Married': ['No', 'Engaged', 'Yes', 'Divorced']})

# perform lookup on df2 using row & column labels from df1
arr = df2.set_index('Answer_Code')\
         .lookup(df1.iloc[:, 1:].values.flatten(),
                 df1.columns[1:].tolist()*3)\
         .reshape(3, -1)

# copy df1 and allocate values from arr
df3 = df1.copy()
df3.iloc[:, 1:] = arr

print(df3)

  Name Graduated Employed    Married
0  AAA       Yes       PT   Divorced
1  BBB        No   Intern        Yes
2  CCC       Yes       No    Engaged

当两个数据帧的列名称匹配时查找值

2 个答案: