Question

我在python pandas中的数据上运行.mean（）但是只返回我正在分组的键和所需列的平均值。我想要所有的列，用平均值替换原始值。尝试了多种东西，但似乎没有任何结果。以下是我用来生成资料的代码

dd1=df.groupby(['key']).agg({'sales':"mean"}).reset_index()

Answer 1

您想要transform ...而不是agg

dd1 = df.groupby('key').transform('mean')

Minimal and Complete Verifiable Example MCVE

df = pd.DataFrame(
    np.random.randn(10, 4)
).add_prefix('C').assign(key=np.random.choice(list('AB'), 10))
print(df)

         C0        C1        C2        C3 key
0 -0.239780  0.167832  0.879349  0.643696   A
1  0.517747  0.573424 -0.480853 -0.162014   A
2  0.236032 -0.396924 -1.406381  1.197946   A
3  0.479451 -0.790073  0.219239 -0.157358   B
4 -0.605864 -0.461622 -1.427521 -1.709760   B
5 -0.281919 -0.965817  1.256316 -1.351529   A
6 -2.085293  0.954725 -1.744391 -1.069667   A
7 -2.100504 -1.161964 -1.102306  0.547207   B
8  1.808283 -0.728799 -1.763971 -1.221539   B
9 -0.975264  0.958484 -0.458139  1.796640   B

现在transform

dd1 = df.groupby('key').transform('mean')
print(dd1)

         C0        C1        C2        C3
0 -0.370642  0.066648 -0.299192 -0.148313
1 -0.370642  0.066648 -0.299192 -0.148313
2 -0.370642  0.066648 -0.299192 -0.148313
3 -0.278780 -0.436795 -0.906540 -0.148962
4 -0.278780 -0.436795 -0.906540 -0.148962
5 -0.370642  0.066648 -0.299192 -0.148313
6 -0.370642  0.066648 -0.299192 -0.148313
7 -0.278780 -0.436795 -0.906540 -0.148962
8 -0.278780 -0.436795 -0.906540 -0.148962
9 -0.278780 -0.436795 -0.906540 -0.148962

您可以通过以下方式将其限制为几列：

cols = ['C0', 'C1', 'C2', 'C3']
dd1 = df.groupby('key')[cols].transform('mean')
print(dd1)

         C0        C1        C2        C3
0 -0.370642  0.066648 -0.299192 -0.148313
1 -0.370642  0.066648 -0.299192 -0.148313
2 -0.370642  0.066648 -0.299192 -0.148313
3 -0.278780 -0.436795 -0.906540 -0.148962
4 -0.278780 -0.436795 -0.906540 -0.148962
5 -0.370642  0.066648 -0.299192 -0.148313
6 -0.370642  0.066648 -0.299192 -0.148313
7 -0.278780 -0.436795 -0.906540 -0.148962
8 -0.278780 -0.436795 -0.906540 -0.148962
9 -0.278780 -0.436795 -0.906540 -0.148962

在python中获取变量的均值

1 个答案: