Pandas 根据另一个 df 上提到的列名称映射两个数据框并部分匹配以派生新列

Question

我有两个数据帧 df1 和 df2 如下。

import pandas as pd


data1 = {'Column1': [1, 2, 3],
        'Column2': ['Account', 'Biscut', 'Super'],
        'Column3': ['Funny', 'Super', 'Nice']}

df1 = pd.DataFrame(data1)

data2 = {'ColumnName':['Column2','Column3'],
     'ifExist':['Acc','Sup'],
     'TarName':['Account_name','Super_name']}
df2 = pd.DataFrame(data2)

我想通过将 df2 中的 ifExists 值与 df2 中提到的 ColumnName 与 df1 部分匹配来将新列

TarName

添加到 df1。

我的预期输出是：

Column1    column2     column3  TarName
1          Account     Funny    Account_Name
2          Biscut      Super    Super_name  
3          Super       Nice 

我尝试过下面的代码。此代码能够部分映射，但只能映射到一列。通过这种方法，我需要创建尽可能多的列映射的字典，并且需要应用尽可能多的列映射。

有更动态的方法吗？

df2_Column2_dict = df2[df2['ColumnName']=='Column2'].set_index(['ifExist'])['TarName'].to_dict()
pat = r'({})'.format('|'.join(df2_Column2_dict.keys()))
extracted = df1['Column2'].str.extract(pat, expand=False).dropna()
df1['TarName'] =  extracted.apply(lambda x: df2_Column2_dict[x]).reindex(df2.index)

print(df1)

Answer 1

我希望对您的代码有所帮助。对于每个概念/行都有注释和解释。

import pandas as pd

# Create DataFrames
data1 = {'Column1': [1, 2, 3],
        'Column2': ['Account', 'Biscut', 'Super'],
        'Column3': ['Funny', 'Super', 'Nice']}
df1 = pd.DataFrame(data1)

data2 = {'ColumnName':['Column2','Column3'],
        'ifExist':['Acc','Sup'],
        'TarName':['Account_name','Super_name']}
df2 = pd.DataFrame(data2)

# Initialize new column in df1
df1['TarName'] = ''

# Iterate over each row in df2
for _, row in df2.iterrows():
    # Column in df1 to search
    column_name = row['ColumnName']  
    
    # Partial value to search for
    if_exist = row['ifExist']        
    
    # Value to assign if there is a match
    tar_name = row['TarName']        

    # Find partial matches and assign TarName
    mask = df1[column_name].str.contains(if_exist, case=False, na=False)
    df1.loc[mask, 'TarName'] = tar_name

# Show the result
print(df1)