Is there a simple way to do an outer join with partially overlapping columns, prioritizing values fr-CodePudding

Suppose I have two dataframes with partially-overlapping indices and partially-overlapping columns -- I want to merge them on the index, keeping both sets of columns but the values of the second DF where they overlap, ie,

old_df = pd.DataFrame({'fruit': ['apples', 'apples', 'bananas', 'bananas'], 
                              'entree': ['steak', 'chicken', 'chicken', 'fish'], 
                              'side': ['fries', 'salad',  'salad', 'soup']}, 
                              index=[0, 1, 2, 3])

new_df = pd.DataFrame({'entree': ['chicken breast', 'salmon', 'ribeye', 'cheeseburger'],                                          
                              'side': ['greek salad', 'clam chowder', 'fries', 'fries'], 
                              'desert': ['key lime pie', 'chocolate mousse', 'tiramasu', np.nan]},                                      
                              index=[2, 3, 4, 5])

merged_df = some_merge_op(old_df, new_df)
merged_df

Desired output:

|   | fruit  | entree         | side         | desert          | 
| - | ------ | -------------- | ------------ | --------------- | 
| 0 | apples | steak          | fries        | nan             |
| 1 | apples | chicken        | salad        | nan             |
| 2 | bananas| chicken breast | greek salad  | key lime pie    |  
| 3 | bananas| salmon         | clam chowder | chocolate mousse| 
| 4 | nan    | ribeye         | fries        | tiramasu        | 
| 5 | nan    | cheeseburger   | fries        | nan             |

CodePudding user response：

You can try combine_first

out = old_df.combine_first(new_df)

print(out)

             desert        entree    fruit   side
0               NaN         steak   apples  fries
1               NaN       chicken   apples  salad
2      key lime pie       chicken  bananas  salad
3  chocolate mousse          fish  bananas   soup
4          tiramasu        ribeye      NaN  fries
5               NaN  cheeseburger      NaN  fries