问题
I have two dataframes
df1 = pd.DataFrame([[1,2],[3,4],[5,6],[7,8]], index = ['a','b','c', 'a'], columns = ['d','e'])
d e
a 1 2
b 3 4
c 5 6
a 7 8
df2 = pd.DataFrame([['a', 10],['b',20],['c',30],['f',40]])
0 1
0 a 10
1 b 20
2 c 30
3 f 40
i want my final dataframe to multiply rows of df1 to multiply by a factor corresponding to value in df2 (for eg. 20 for b)
so my output should look like
d e
a 10 20
b 60 80
c 150 180
a 70 80
Kindly provide a solution assuming df1 to be hundreds of rows in length. I could only think of looping through df1.index.
回答1:
Use set_index and reindex to align df2 with df1 and then mul
In [1150]: df1.mul(df2.set_index(0).reindex(df1.index)[1], axis=0)
Out[1150]:
d e
a 10 20
b 60 80
c 150 180
a 70 80
回答2:
Create a mapping and call df.apply:
In [1128]: mapping = dict(df2.values)
In [1129]: df1.apply(lambda x: x * mapping[x.name], 1)
Out[1129]:
d e
a 10 20
b 60 80
c 150 180
a 70 80
回答3:
IIUC:
In [55]: df1 * pd.DataFrame(np.tile(df2[[1]],2), columns=df1.columns, index=df2[0])
Out[55]:
d e
a 10 20
a 70 80
b 60 80
c 150 180
Helper DF:
In [57]: pd.DataFrame(np.tile(df2[[1]],2), columns=df1.columns, index=df2[0])
Out[57]:
d e
0
a 10 10
b 20 20
c 30 30
回答4:
This is straight forward. You just make sure they have a common axis, then you can combine them:
put the lookup column into the index
df2.set_index(0, inplace=True)
1
0
a 10
b 20
c 30
Now you can put that column into df1 very easily:
df1['multiplying_factor'] = df2[1]
Now you just want to multiply two columns:
df1['final_value'] = df1.e*df1.multiplying_factor
Now df1 looks like:
d e multiplying_factor final_value
a 1 2 10 20
b 3 4 20 80
c 5 6 30 180
a 7 8 10 80
来源：https://stackoverflow.com/questions/45801885/multiply-dataframe-with-values-from-other-dataframe