0

df1：

id    Stu_Name   Class     Fees
1     Jack       primary   2333
2     mack       primary   2363
3     may        primary   2833
3     Mark       primary   1333
3     John       primary   9333
4     Moon       Secondary 6589 
5     daisy      Secondary 6565 
6     shawn      Secondary 6545 
6     roy        Secondary 6596
9     hary       higher    8526
10    Joy        higher    9654
10    nick       higher    7845
10    julie      higher    9633

df2:

id    Stu_Name   Class     Fees
11    eric       primary   2333
21    fick       primary   2363
42    Moon       Secondary 6589 
56    anki       Secondary 6565 
18    menk       higher    7845
17    rock       higher    9633

df_输出：

id    Stu_Name   Class     Fees
1     Jack       primary   2333
2     mack       primary   2363
11    eric       primary   2333
21    fick       primary   2363
3     may        primary   2833
3     Mark       primary   1333
3     John       primary   9333
4     Moon       Secondary 6589 
5     daisy      Secondary 6565 
42    Moon       Secondary 6589 
56    anki       Secondary 6565
6     shawn      Secondary 6545 
6     roy        Secondary 6596
9     hary       higher    8526
18    menk       higher    7845
17    rock       higher    9633
10    Joy        higher    9654
10    nick       higher    7845
10    julie      higher    9633

我需要将两个 df 与行放置相连接，例如：您可以在 df1 中看到类“primary”Id“3”重复了几次，我需要在 id 之前附加 df2 类“primary”行df1 正在重复。其他班级也是如此。

4

1 回答 1

1

我误解了目标。这是修改后的回复。这是一个很长的响应，但工作是在三行代码中完成的。

首先，创建并合并两个数据框：

from io import StringIO
import pandas as pd

# create the 2 data frames

data = '''id    Stu_Name   Class     Fees
1     Jack       primary   2333
2     mack       primary   2363
3     may        primary   2833
3     Mark       primary   1333
3     John       primary   9333
4     Moon       Secondary 6589 
5     daisy      Secondary 6565 
6     shawn      Secondary 6545 
6     roy        Secondary 6596
9     hary       higher    8526
10    Joy        higher    9654
10    nick       higher    7845
10    julie      higher    9633
'''
df1 = pd.read_csv(StringIO(data), sep='\s+', engine='python')


data = '''id    Stu_Name   Class     Fees
11    eric       primary   2333
21    fick       primary   2363
42    Moon       Secondary 6589 
56    anki       Secondary 6565 
18    menk       higher    7845
17    rock       higher    9633
'''
df2 = pd.read_csv(StringIO(data), sep='\s+', engine='python')

# combine the 2 data frames
df = pd.concat([df1, df2], ignore_index=True)

现在，创建两个帮助列，然后排序：

# create the 1st helper column (for sorting at end)
# this will group (and sort) primary, Secondary, higher
df['class_num'] = df['Class'].factorize()[0]

# create 2nd helper column (to identify repeated IDs)
df['id_count'] = df.groupby('id')['id'].transform('count')

# if the logic is correct, then drop 'class_num', 'id_count'
df = df.sort_values(['class_num', 'id_count']).set_index('id')

结果是：

print(df)

   Stu_Name      Class  Fees  class_num  id_count
id                                               
1      Jack    primary  2333          0         1
2      mack    primary  2363          0         1
11     eric    primary  2333          0         1
21     fick    primary  2363          0         1
3       may    primary  2833          0         3
3      Mark    primary  1333          0         3
3      John    primary  9333          0         3
4      Moon  Secondary  6589          1         1
5     daisy  Secondary  6565          1         1
42     Moon  Secondary  6589          1         1
56     anki  Secondary  6565          1         1
6     shawn  Secondary  6545          1         2
6       roy  Secondary  6596          1         2
9      hary     higher  8526          2         1
18     menk     higher  7845          2         1
17     rock     higher  9633          2         1
10      Joy     higher  9654          2         3
10     nick     higher  7845          2         3
10    julie     higher  9633          2         3

原帖

您可以使用 Categorical 类型来提供自定义排序顺序：

class_idx = pd.CategoricalIndex(
    categories = ['primary', 'Secondary', 'higher'], 
    ordered=True)

df = pd.concat([df1, df2]).astype(
    {'id': 'int', 
     'Stu_Name': 'string', 
     'Class': class_idx, 
     'Fees': 'int'}).sort_values('Class')

print(df.head())

   id Stu_Name    Class  Fees
0   1     Jack  primary  2333
1   2     mack  primary  2363
2   3      may  primary  2833
3   3     Mark  primary  1333
4   3     John  primary  9333
于 2020-08-07T14:31:32.243 回答