访问嵌套字典中的单元格的最佳方式是什么？

Question

我有一本嵌套的字典。它从(STRING, DATE)元组到包含特定列和值的字典。我一直在试图找出获取单个单元格数据的语法。例如，我想打电话给[（'SYY'，'1997-06-30'）]['dvt']，得到99.5740。我试过用数据框获取和[（元组）]的各种形式，以及没有任何运气。我收到的错误是

KeyError: ('SYY', '2016-06-30')

我拥有的熊猫在印刷时是这样的：

                   ivncf       dvt       fincf       chech  
SYY 1997-06-30  -213.3560   99.5740   -274.8150      9.9370  
    1998-06-30  -335.5300  110.9280    -29.6420     -7.4080  
    1999-06-30  -261.7350  126.6910   -284.5530     39.0150  
    2000-06-30  -459.3920  145.4180   -239.5090      9.8250  
    2001-06-30  -338.7510  173.7010   -639.8580    -23.3850  
    2002-06-30  -630.3000  225.5300   -359.9840     94.6960  
    2003-06-30  -681.8250  273.8520   -550.5280    139.0080  
    2004-06-30  -683.8110  321.3530   -641.8510   -137.7410  
    2005-06-30  -413.4400  368.7920   -784.2700     -8.0280  
    2006-06-30  -608.6710  408.2640   -504.6170     10.2190  
    2007-06-30  -648.7110  456.4380   -748.2500      5.9750  
    2008-06-30  -555.5600  513.5930   -698.5780    343.6800  
    2009-06-30  -658.6630  557.4870   -379.6430    535.5320  
    2010-06-30  -656.3200  585.7340   -667.0300   -433.2080  
    2011-06-30  -679.5560  604.5000   -377.9070     54.3220  
    2012-06-30  -903.6290  628.0240   -442.6490     49.1020  
    2013-06-30  -911.8820  654.8710   -874.2080   -276.5820  
    2014-06-30  -576.8380  673.5680   -915.8580      0.7610  
    2015-06-30  -654.3460  705.5390   3897.5620   4716.9980  
    2016-06-30  -600.8280  695.4690  -2404.7310  -1210.7440 

有什么建议吗？谢谢！你知道吗

Answer 1

我想你需要^{}：

print (df.loc[('SYY', '1997-06-30'), 'dvt'])
99.574

对于复杂的选择，请使用slicers：

idx = pd.IndexSlice
print (df.loc[idx['SYY', '1997-06-30'], 'dvt'])
99.574

---

idx = pd.IndexSlice
print (df.loc[idx['SYY', '1997-06-30':'2001-06-30'], idx['dvt':'fincf']])
                    dvt    fincf
SYY 1997-06-30   99.574 -274.815
    1998-06-30  110.928  -29.642
    1999-06-30  126.691 -284.553
    2000-06-30  145.418 -239.509
    2001-06-30  173.701 -639.858

---

idx = pd.IndexSlice
print (df.loc[idx['SYY', '1997-06-30'], idx['dvt':'chech']])
dvt       99.574
fincf   -274.815
chech      9.937
Name: (SYY, 1997-06-30), dtype: float64

---

idx = pd.IndexSlice
print (df.loc[idx['SYY', '1997-06-30':'2003-06-30'], 'dvt'])
SYY  1997-06-30     99.574
     1998-06-30    110.928
     1999-06-30    126.691
     2000-06-30    145.418
     2001-06-30    173.701
     2002-06-30    225.530
     2003-06-30    273.852
Name: dvt, dtype: float64