pandas - 使用 groupby 和条件对 Pandas 数据框进行排序

Question

我正在尝试根据满足条件的组对数据框进行排序。

我在对组进行排序时遇到语法错误。在尝试上述操作之前，我丢失了数据框的初始顺序。

这是我要实现的排序顺序:

1) 按第一列和测试列排序。

2) Test==1 groups, sort on Secondary and then by Final column.

---测试==0组，仅在最后一列排序。

import pandas as pd

df=pd.DataFrame({"First":[100,100,100,100,100,100,200,200,200,200,200],"Test":[1,1,1,0,0,0,0,1,1,1,0],"Secondary":[.1,.1,.1,.2,.2,.3,.3,.3,.3,.4,.4],"Final":[1.1,2.2,3.3,4.4,5.5,6.6,7.7,8.8,9.9,10.10,11.11]})

def sorter(x):
    if x["Test"]==1:
        x.sort_values(['Secondary','Final'], inplace=True)
    else:
        x=x.sort_values('Final', inplace=True)


df=df.sort_values(["First","Test"],ascending=[False, False]).reset_index(drop=True)


df.groupby(['First','Test']).apply(lambda x: sorter(x))

df

---

Expected result:

First Test Secondary Final
200     1   0.4     10.1
200     1   0.3*    9.9*
200     1   0.3*    8.8*

200     0   0.4     11.11*
200     0   0.3     7.7*

100     1   0.5     2.2
100     1   0.1*    3.3*
100     1   0.1*    1.1*

100     0   0.3     6.6*
100     0   0.2     5.5*
100     0   0.2     4.4*

Answer 1

你可以试试不用groupby按降序排序，w.r.t 你给的顺序，排序顺序会改变。它对你有用吗<​​/p>

df=pd.DataFrame({"First":[100,100,100,100,100,100,200,200,200,200,200],"Test":[1,1,1,0,0,0,0,1,1,1,0],"Secondary":[.1,.5,.1,.9,.4,.1,.3,.3,.3,.4,.4],"Final":[1.1,2.2,3.3,4.4,5.5,6.6,7.7,8.8,9.9,10.10,11.11]})

df = df.groupby(['First','Test']).apply(lambda x: x.sort_values(['First','Test','Secondary','Final'],ascending=False) if x.iloc[0]['Test']==1 else x.sort_values(['First','Test','Final'],ascending=False)).reset_index(drop=True)
df.sort_values(['First','Test'],ascending=[True,False])

输出:

    Final   First   Secondary   Test
3   2.20    100 0.5 1
4   3.30    100 0.1 1
5   1.10    100 0.1 1
0   6.60    100 0.1 0
1   5.50    100 0.4 0
2   4.40    100 0.9 0
8   10.10   200 0.4 1
9   9.90    200 0.3 1
10  8.80    200 0.3 1
6   11.11   200 0.4 0
7   7.70    200 0.3 0