This is something that I would be able to replicate easily in Excel with an XLOOKUP function, but I'm trying to do it with pandas.
I have 2 dataframes, say something like this:
df1
|first_name | last_name | dob | value |
| Goku | Saiyan | 1/1/2021 | |
| Vegetta | Super | 8/7/1990 | |
| Gohan | Son | 4/20/1969| |
df2
|first_name | last_name | dob | value |
| Goku | Saiyan | 1/1/2021 | 50 |
| Vegetta | Super | 8/7/1990 | 92 |
| Gohan | Son | 4/20/1969| 31 |
| Trunks | Donald | 7/1/1921 | 49 |
| New Name | Another | 1/31/1992| 67 |
I would like to fill the value
column in df1 from the value
column in df2.
I cannot use combine_first
because the dataframes have different index and different sizes.
If I use pd.merge
then I get the value_x
and value_y
where value_y
has the data that I want, but I need to do more process to have it where I want on df1['value']
I basically want to match the first name, last name and dob on both dataframes and receive the value
from df2.
It's probably a simple issue, but I have been struggling with the different methods that I've tried and I think there must be something that I'm missing because it shouldn't be that complicated.
Any help will be really appreciated.
CodePudding user response:
If your value
column from df1
does not contain existing value, you can drop it and use merge
:
>>> pd.merge(df1.drop(columns='value'), df2, how='left',
on=['first_name', 'last_name', 'dob'])
first_name last_name dob value
0 Goku Saiyan 1/1/2021 50
1 Vegetta Super 8/7/1990 92
2 Gohan Son 4/20/1969 31
CodePudding user response:
Use map.
Create a dict of firts_name; value from df2 and map to df1's first_name.
df1 =df1.assign(value=df1['first_name'].map(dict(zip(df2['first_name'],df2['value']))))
first_name last_name dob value
0 Goku Saiyan 1/1/2021 50
1 Vegetta Super 8/7/1990 92
2 Gohan Son 4/20/1969 31