使用 pandas 數據幀映射到中轉（節點）點的跨容量請求

Question

輸入帶有 TerminalID、TName、XY 坐標、PeopleID 的數據框

import pandas as pd

data = {
        'TerminalID': ['5','21','21','2','21','2','5','22','22','22','2','32','41','41','42','50','50'],
        'TName': ['AD','AMBO','AMBO','PS','AMBO','PS','AD','AM','AM','AM','PS','BO','BA','BA','BB','AZ','AZ'],
        'xy': ['1.12731,1.153756','0.12731,0.153757','0.12731,0.153757','1.989385,1.201941','0.12731,0.153757','1.989385,1.201941','1.12731,1.153756','2.12731,1.153756','2.12731,1.153756','2.12731,1.153756','1.989385,1.201941','1.989385,1.201941','2.989385,1.201941','2.989385,1.201941','2.989385,3.201941','3.989385,3.201941','3.989385,3.201941'],
        'Pcode': [ 'None','Z014','Z015','Z016','Z017','Z018','None','Z020','Z021','Z022','Z023','Z024','Z025','Z026','Z027','Z028','Z029']
    }

df = pd.DataFrame.from_dict(data)

出[55]：

DF1的輸出

   TerminalID TName                 xy Pcode
0           5    AD   1.12731,1.153756  None
1          21  AMBO   0.12731,0.153757  Z014
2          21  AMBO   0.12731,0.153757  Z015
3           2    PS  1.989385,1.201941  Z016
4          21  AMBO   0.12731,0.153757  Z017
5           2    PS  1.989385,1.201941  Z018
6           5    AD   1.12731,1.153756  None
7          22    AM   2.12731,1.153756  Z020
8          22    AM   2.12731,1.153756  Z021
9          22    AM   2.12731,1.153756  Z022
10          2    PS  1.989385,1.201941  Z023
11         32    BO  1.989385,1.201941  Z024
12         41    BA  2.989385,1.201941  Z025
13         41    BA  2.989385,1.201941  Z026
14         42    BB  2.989385,3.201941  Z027
15         50    AZ  3.989385,3.201941  Z028
16         50    AZ  3.989385,3.201941  Z029

DF2,

T_cap 是終端 ID 的容量需求，T_load 是負載細節，Tcap 是運行計數增量，T_load 是終端的實際請求，開始和結束的 0 是解決方案的填充

data2= {
        'BusID': ['18','18','18','18','18','18','18','18','18'],
        'Tcap': ['0','2','3','6','7','8','10','12','12'],
        'T_Load': ['0','2','1','2','2','1','2','2','0'],
        'TerminalID': [ '5','21','33','2','32','42','41','50','5'],
        
        'TName':['AD','AMBO','AM','PS','BO','BB','BA','AZ','AD']
    }

df2 = pd.DataFrame.from_dict(data2)

出[59]：

  BusID Tcap T_Load TerminalID TName
0    18    0      0          5    AD
1    18    2      2         21  AMBO
2    18    3      1         33    AM
3    18    6      2          2    PS
4    18    7      2         32    BO
5    18    8      1         42    BB
6    18   10      2         41    BA
7    18   12      2         50    AZ
8    18   12      0          5    AD
    

Data Frame # 請求的最終輸出

輸出基於 T_Load 約束。

data3 = {
        'BusID': ['18','18','18','18','18','18','18','18','18'],
        'Tcap': ['0','2','3','6','7','8','10','12','12'],
        'T_Load': ['0','2','1','3','1','1','2','2','0'],
        'TerminalID': [ '5','21','33','2','32','42','41','50','5'],
        
        'TName':['AD','AMBO','AM','PS','BO','BB','BA','AZ','AD'],
        'Pcode':['None','Z013,Z019','Z020','Z016,Z018,Z023','Z024','Z027','Z025,Z026','Z028,Z029','None']
    }
    
    df3 = pd.DataFrame.from_dict(data3)

出[61]：

  BusID Tcap T_Load TerminalID TName           Pcode
0    18    0      0          5    AD            None
1    18    2      2         21  AMBO       Z013,Z019
2    18    3      1         33    AM            Z020
3    18    6      3          2    PS  Z016,Z018,Z023
4    18    7      1         32    BO            Z024
5    18    8      1         42    BB            Z027
6    18   10      2         41    BA       Z025,Z026
7    18   12      2         50    AZ       Z028,Z029
8    18   12      0          5    AD            None

感謝您

Answer 1

我的解決方案按TerminalID和TName聚合連接，並通過聚合列表分配給另一個 DataFrame，最后通過join列表理解中的位置過濾值：

s = df.groupby(['TerminalID','TName'])['Pcode'].agg(list).rename('P_list')
df = df2.join(s, on=['TerminalID','TName'])

df['P_list'] = [','.join(x[:int(y)]) if int(y) != 0 else None 
                for x, y in zip(df['P_list'], df['T_Load'])]
print (df)
  BusID Tcap T_Load TerminalID TName          P_list
0    18    0      0          5    AD            None
1    18    2      2         21  AMBO       Z014,Z015
2    18    3      1         22    AM            Z020
3    18    6      3          2    PS  Z016,Z018,Z023
4    18    7      1         32    BO            Z024
5    18    8      1         42    BB            Z027
6    18   10      2         41    BA       Z025,Z026
7    18   12      2         50    AZ       Z028,Z029
8    18   12      0          5    AD            None

Answer 2

您可以map每個 TName 的聚合字符串：

df2['Plist'] = df2['TName'].map(df.groupby('TName')['Pcode'].agg(','.join))

或者，如果要將多個字符串 None 替換為單個字符串：

df2['Plist'] = df2['TName'].map(df.groupby('TName')['Pcode']
                                  .agg(lambda x: ','.join(e for e in x if e != 'None'))
                                  .replace('', 'None')
                                )

輸出：

  BusID Tcap T_Load TerminalID TName           Plist
0    18    0      0          5    AD            None
1    18    2      2         21  AMBO  Z014,Z015,Z017
2    18    3      1         22    AM  Z020,Z021,Z022
3    18    6      3          2    PS  Z016,Z018,Z023
4    18    7      1         32    BO            Z024
5    18    8      1         42    BB            Z027
6    18   10      2         41    BA       Z025,Z026
7    18   12      2         50    AZ       Z028,Z029
8    18   12      0          5    AD            None

更新：限制輸出：

然后，您可以使用正則表達式修剪列，我們可以使用groupby從每個組中的矢量化字符串操作中受益（如果組少而行多，這最有趣）：

df2['P_list'] = (df2.groupby('T_Load')['P_list']
                    .apply(lambda c: c.str.extract(rf'((?:[^,]+,?){{,{str(c.name)}}})',
                                                   expand=False)
                                    .str.strip(',')
                          )
                    .replace('', 'None')
                )

輸出：

  BusID Tcap T_Load TerminalID TName          P_list
0    18    0      0          5    AD            None
1    18    2      2         21  AMBO       Z014,Z015
2    18    3      1         22    AM            Z020
3    18    6      3          2    PS  Z016,Z018,Z023
4    18    7      1         32    BO            Z024
5    18    8      1         42    BB            Z027
6    18   10      2         41    BA       Z025,Z026
7    18   12      2         50    AZ       Z028,Z029
8    18   12      0          5    AD            None