Question

我有2个数据框。如何根据特定的列将2个不相等的数据帧相乘，是否有任何pandas包或需要使用循环？有人可以共享将两个不等维的数据帧相乘的代码。

df1 is below:

    features   tf_vals
0       this  0.200000
1         is  0.200000
2        the  0.200000
3      first  0.200000
4   document  0.200000
5       this  0.166667
6   document  0.333333
7         is  0.166667
8        the  0.166667
9     second  0.166667
10       and  0.166667
11      this  0.166667


df2 is :

   features  idf_vals
0       the  1.000000
1      this  1.000000
2  document  1.000000
3        is  1.000000
4     first  1.510826
5       one  1.916291
6    second  1.916291
7       and  1.916291
8     third  1.916291

How do I multiply df1 and df2 based on column name 'features'.

Desired output:
I want to multiply value of word in df1 and value of word in df2. example(value of word 'this' in df1 * value of word 'this' in df2 )

features   Mult
this       0.2
is         0.2
the        0.2
first      0.3021652
document   0.2
this       0.166667
document   0.333333
is         0.166667
the        0.166667
second     0.319382472
and        0.319382472
this       0.166667

Answer 1

为避免对df1['features']列进行排序，想法由features使用Series.map，然后由Series.mul使用倍数：

s = df2.set_index('features')['idf_vals']
df1['Mult'] = df1['features'].map(s).mul(df1.pop('tf_vals'))

print (df1)
    features      Mult
0       this  0.200000
1         is  0.200000
2        the  0.200000
3      first  0.302165
4   document  0.200000
5       this  0.166667
6   document  0.333333
7         is  0.166667
8        the  0.166667
9     second  0.319382
10       and  0.319382
11      this  0.166667

如何基于列值乘以2个不相等的数据帧？

1 个答案: