重複IFERROR和VLOOKUP在大熊貓加入

Question

我想連接兩個dataframes：

df1 = pd.DataFrame({'Banner': {0: 'banner1', 1: 'banner2', 2: 'banner3'}, 
        'Campaign': {0: 'campaign1', 1: 'campaign2', 2: '12345'}, 
        'Country ': {0: 'de', 1: 'it', 2: 'de'}, 
        'Date': {0: '1/1/2016', 1: '2/1/2016', 2: '1/1/2016'}, 
        'Value_1': {0: 10, 1: 5, 2: 20}}) 

df2 = pd.DataFrame({'Banner': {0: 'banner1', 1: 'banner2', 2: 'banner3', 3: 'banner4', 4: 'banner5'}, 
        'Campaign': {0: 'campaign1',1: 'campaign2', 2: 'none',3: 'campaign4',4: 'campaign5'}, 
        'Country ': {0: 'de', 1: 'it', 2: 'de', 3: 'en', 4: 'en'}, 
        'Date': {0: '1/1/2016', 1: '2/1/2016', 2: '1/1/2016', 3: '3/1/2016', 4: '4/1/2016'}, 
        'Value_2': {0: 5, 1: 10, 2: 15, 3: 20, 4: 25},    
        'id_campaign': {0: 'none', 1: 'none', 2: '12345', 3: 'none', 4: 'none'}}) 

編輯： 讓我們想象的選項：

df1 = pd.DataFrame({'Banner': {0: 'banner1', 1: 'banner2', 2: 'banner3'}, 
       'Campaign': {0: 'campaign1', 1: 'campaign2', 2: '12345'}, 
       'Date': {0: '1/1/2016', 1: '2/1/2016', 2: '1/1/2016'}, 
       'Value_1': {0: 10, 1: 5, 2: 20}}) 

我不得不加入DF2與DF1在鍵上：

• 日期
• 運動
• 橫幅

這裏的問題是，當在關鍵「戰役」中沒有找到匹配的，關鍵應該切換到字段「id_campaign」。

我想獲得該數據幀：

df_joined = pd.DataFrame({'Banner': {0: 'banner1', 1: 'banner2', 2: 'banner3', 3: 'banner4', 4: 'banner5'}, 
         'Campaign': {0: 'campaign1', 1: 'campaign2', 2: 'none', 3: 'campaign4', 4: 'campaign5'}, 
         'Country ': {0: 'de', 1: 'it', 2: 'de', 3: 'en', 4: 'en'}, 
         'Date': {0: '1/1/2016', 1: '2/1/2016', 2: '1/1/2016', 3: '3/1/2016', 4: '4/1/2016'}, 
         'Value_1': {0: 10, 1: 5, 2: 20, 3: 0, 4: 0}, 
         'Value_2': {0: 5, 1: 10, 2: 15, 3: 20, 4: 25}, 
         'id_campaign': {0: 'none', 1: 'none', 2: '12345', 3: 'none', 4: 'none'}}) 

任何幫助非常感謝。

Answer 1

您可以使用雙merge 3和2鍵，然後通過combine_first從df4Value_1欄填不匹配值：

df3 = pd.merge(df2, df1.drop('Country', axis=1), on=['Date','Campaign','Banner'], how='left') 
df4 = pd.merge(df2, df1, on=['Date','Banner'], how='left') 

print (df3) 
    Banner Campaign Country  Date Value_2 id_campaign Value_1 
0 banner1 campaign1  de 1/1/2016  5  none  10.0 
1 banner2 campaign2  it 2/1/2016  10  none  5.0 
2 banner3  none  de 1/1/2016  15  12345  NaN 
3 banner4 campaign4  en 3/1/2016  20  none  NaN 
4 banner5 campaign5  en 4/1/2016  25  none  NaN 

print (df4['Value_1']) 
0 10.0 
1  5.0 
2 20.0 
3  NaN 
4  NaN 
Name: Value_1, dtype: float64 

df3['Value_1'] = df3['Value_1'].combine_first(df4['Value_1']).fillna(0).astype(int) 
print (df3) 
    Banner Campaign Country  Date Value_2 id_campaign Value_1 
0 banner1 campaign1  de 1/1/2016  5  none  10 
1 banner2 campaign2  it 2/1/2016  10  none  5 
2 banner3  none  de 1/1/2016  15  12345  20 
3 banner4 campaign4  en 3/1/2016  20  none  0 
4 banner5 campaign5  en 4/1/2016  25  none  0