如果一個DataFrame的index亂了，按正常的sort_index無法排序，則可以用reindex函式

import numpy as np
import pandas as pd
>>> randWeek=['Tues','Thur','Mon','Sat','Fri','Sun','Wed']
>>> frame=pd.DataFrame(np.random.randn(7,3),columns=list('abc'),index=randWeek)
>>> frame
                 a         b         c
    Tues  0.054086
 
 -2.546081  0.051492
    Thur  0.232368 -0.937215  0.883313
    Mon  -1.004652  1.266954  0.321374
    Sat   2.003404  0.514108  0.640477
    Fri  -0.631527  0.939361 -0.013297
    Sun   0.492959 -0.171945  0.030827
    Wed  -0.334522 -0.334029  0.575805

>>> normalWeek=['Mon','Tues','Wed','Thur','Fri','Sat','Sun']
>>> 
 
frame.reindex(normalWeek)
                 a         b         c
    Mon  -1.004652  1.266954  0.321374
    Tues  0.054086 -2.546081  0.051492
    Wed  -0.334522 -0.334029  0.575805
    Thur  0.232368 -0.937215  0.883313
    Fri  -0.631527  0.939361 -0.013297
    Sat   2.003404  0.514108  0.640477
    Sun   0.492959 -0.171945  0.030827
 

指定的list中的index可以比DataFrame中已有的index多或少

>>> otherWeek1=['Tues','Wed','Thur','Fri','Sat','Sun']
>>> frame.reindex(otherWeek1)
                 a         b         c
    Tues  0.054086 -2.546081  0.051492
    Wed  -0.334522 -0.334029  0.575805
    Thur  0.232368 -0.937215  0.883313
    Fri  -0.631527  0.939361 -0.013297
    Sat   2.003404  0.514108  0.640477
    Sun   0.492959 -0.171945  0.030827

>>> otherWeek2=['Mon','Tues','Wed','Thur','Fri','Sat','Sun','lalala']
>>> frame.reindex(otherWeek2)
                   a         b         c
    Mon    -1.004652  1.266954  0.321374
    Tues    0.054086 -2.546081  0.051492
    Wed    -0.334522 -0.334029  0.575805
    Thur    0.232368 -0.937215  0.883313
    Fri    -0.631527  0.939361 -0.013297
    Sat     2.003404  0.514108  0.640477
    Sun     0.492959 -0.171945  0.030827
    lalala       NaN       NaN       NaN

人生苦短，我用python~