使用不同的篩選將dataframe自身連接起來

我有一個(gè)df遵循這個(gè)結(jié)構(gòu)：

store       day               type              sales       orders

amazon      2021-10-10        web               10          1
amazon      2021-10-10        retail            500         50
facebook    2021-10-10        retail            300         50
facebook    2021-09-05        retail            10          50
apple       2021-09-01        web               5           1
uber        2021-08-01        web               50          1
uber        2021-08-01        retail            60          1
...

我正在嘗試一個(gè)df_res每周sales & orders次的store, day & type，因此輸出結(jié)果如下所示：

day             type      sales_amazon        orders_amazon       sales_facebook       orders_facebook     sales_apple        orders_apple        sales_uber      orders_uber
2021-08-01      web       0                   0                   0                    0                   0                  0                   50              1     
2021-08-01      rtail     0                   0                   0                    0                   0                  0                   60              1    
2021-10-10      web       10                  1                   0                    0                   0                  0                   0               0      
2021-10-10      retail    500                 50                  300                  50                  0                  0                   0               0 
...   

I tried:

# main df to be joined on
df_res = df[df.store.isin(['amazon'])].groupby(['store','type', pd.Grouper(key = 'day', freq = 'W-MON', 
                                               label = 'right')])[['store','day','orders','sales','type']].sum().reset_index()

# merging on main df each store df
for branch in ['facebook', 'apple', 'apple', 'uber']:
    df_res = df_res.merge(df[df.store.isin([branch])].groupby(['store','type', pd.Grouper(key = 'day', freq = 'W-MON',
                          label = 'right')])[['store','day','orders','sales','type']].sum().reset_index(), 
                          on =['day','type'], suffixes= [f'_{branch}', f'_{branch}'], how = 'outer')

但這并沒有產(chǎn)生我想要的結(jié)構(gòu)，我嘗試使用join，但這會(huì)拋出一個(gè)不同的長(zhǎng)度錯(cuò)誤，因?yàn)橛袝r(shí)給定商店的特定date & type組合沒有銷售。