給定索引列表，如何在這些索引上拆分DataFrame？

Question

給定一個任意長度的索引列表，例如[6, 12]，在這些索引上拆分DataFrame的最佳方法是什麼，所以最後我有三個不同的DataFrame？

Answer 1

這是處理非整數索引的更一般的策略。

import pandas as pd 
import numpy as np 

np.random.seed(0) 
df = pd.DataFrame(np.random.randn(100, 2), columns=['A', 'B']).set_index('A') 

Out[32]: 
       B 
A    
1.7641 0.4002 
0.9787 2.2409 
1.8676 -0.9773 
0.9501 -0.1514 
-0.1032 0.4106 
0.1440 1.4543 
0.7610 0.1217 
0.4439 0.3337 
1.4941 -0.2052 
0.3131 -0.8541 
-2.5530 0.6536 
0.8644 -0.7422 
2.2698 -1.4544 
0.0458 -0.1872 
1.5328 1.4694 
...   ... 
0.9209 0.3187 
0.8568 -0.6510 
-1.0342 0.6816 
-0.8034 -0.6895 
-0.4555 0.0175 
-0.3540 -1.3750 
-0.6436 -2.2234 
0.6252 -1.6021 
-1.1044 0.0522 
-0.7396 1.5430 
-1.2929 0.2671 
-0.0393 -1.1681 
0.5233 -0.1715 
0.7718 0.8235 
2.1632 1.3365 

[100 rows x 1 columns] 

# say you want -1 and 1 as split point 
cutoffs = np.array([-np.inf, -1, 1, np.inf]) 
df['group_level'] = pd.cut(df.index.values, bins=cutoffs) 
# then groupby will give you what you want. 
grouped = df.groupby('group_level') 

for group_name, group_df in grouped: 
    print(group_name) 
    print(group_df) 

(-inf, -1] 
       B group_level 
A       
-2.5530 0.6536 (-inf, -1] 
-1.0486 -1.4200 (-inf, -1] 
-1.7063 1.9508 (-inf, -1] 
-1.2528 0.7775 (-inf, -1] 
-1.6139 -0.2127 (-inf, -1] 
-1.6302 0.4628 (-inf, -1] 
-1.1651 0.9008 (-inf, -1] 
-1.0708 1.0545 (-inf, -1] 
-1.2705 0.9694 (-inf, -1] 
-1.1731 1.9436 (-inf, -1] 
-1.4913 0.4394 (-inf, -1] 
-1.3159 -0.4616 (-inf, -1] 
-1.1475 -0.4378 (-inf, -1] 
-1.2254 0.8444 (-inf, -1] 
-1.0002 -1.5448 (-inf, -1] 
-1.0342 0.6816 (-inf, -1] 
-1.1044 0.0522 (-inf, -1] 
-1.2929 0.2671 (-inf, -1] 
(-1, 1] 
       B group_level 
A       
0.9787 2.2409  (-1, 1] 
0.9501 -0.1514  (-1, 1] 
-0.1032 0.4106  (-1, 1] 
0.1440 1.4543  (-1, 1] 
0.7610 0.1217  (-1, 1] 
0.4439 0.3337  (-1, 1] 
0.3131 -0.8541  (-1, 1] 
0.8644 -0.7422  (-1, 1] 
0.0458 -0.1872  (-1, 1] 
0.1549 0.3782  (-1, 1] 
-0.8878 -1.9808  (-1, 1] 
-0.3479 0.1563  (-1, 1] 
-0.3873 -0.3023  (-1, 1] 
-0.5097 -0.4381  (-1, 1] 
-0.8955 0.3869  (-1, 1] 
...   ...   ... 
-0.7448 -0.8264  (-1, 1] 
-0.0985 -0.6635  (-1, 1] 
-0.4980 1.9295  (-1, 1] 
0.9494 0.0876  (-1, 1] 
0.9209 0.3187  (-1, 1] 
0.8568 -0.6510  (-1, 1] 
-0.8034 -0.6895  (-1, 1] 
-0.4555 0.0175  (-1, 1] 
-0.3540 -1.3750  (-1, 1] 
-0.6436 -2.2234  (-1, 1] 
0.6252 -1.6021  (-1, 1] 
-0.7396 1.5430  (-1, 1] 
-0.0393 -1.1681  (-1, 1] 
0.5233 -0.1715  (-1, 1] 
0.7718 0.8235  (-1, 1] 

[66 rows x 2 columns] 
(1, inf] 
      B group_level 
A       
1.7641 0.4002 (1, inf] 
1.8676 -0.9773 (1, inf] 
1.4941 -0.2052 (1, inf] 
2.2698 -1.4544 (1, inf] 
1.5328 1.4694 (1, inf] 
1.2303 1.2024 (1, inf] 
1.1394 -1.2348 (1, inf] 
1.4883 1.8959 (1, inf] 
1.1788 -0.1799 (1, inf] 
1.8832 -1.3478 (1, inf] 
1.9229 1.4805 (1, inf] 
1.8676 0.9060 (1, inf] 
2.3831 0.9445 (1, inf] 
1.1266 -1.0799 (1, inf] 
1.1880 0.3169 (1, inf] 
2.1632 1.3365 (1, inf]