如何處理這個與dataframe.sort相關的熊貓錯誤？

Question

一個數據幀「df，可以」：

>>> df = ACEV(get_SW_code('sdht', sw_cls=3))[['PB','PE','EV_EBITDA','PEG','ROIC','mg_r','opr_pft_r','net_pft_r','sales_gr','net_pft_r','Ttl_mkv']] 
>>> print df 
         PB  PE EV_EBITDA  PEG ROIC mg_r opr_pft_r net_pft_r sales_gr net_pft_r Ttl_mkv 
STK_ID RPT_Date                            
000153 20130930 2.4196 96.1577 28.4962 2.4451 0.0215 0.2420  0.0169  0.0205 -0.1744  0.0205 24.4094 
000513 20130930 3.6926 25.0706 18.4354 2.4541 0.1031 0.6270  0.1375  0.1171 0.1115  0.1171 119.5308 
000566 20130930 3.4557 71.1281 48.0766 -10..0624 0.3442  0.1013  0.1048 0.2624  0.1048 50.3113 
000915 20130930 7.6417 36.7593 15.4284 1.0311 0.2811 0.6573  0.3108  0.1501 0.3122  0.1501 55.6988 
002262 20130930 10.8237 51.6500 35.6455 2.0330 0.2355 0.3906  0.1031  0.0883 0.1175  0.0883 87.4037 
002294 20130930 7.2780 26.5449 21.4806 0.7326 0.2700 0.7687  0.4328  0.3649 0.2072  0.3649 210.4453 
002332 20130930 3.8046 56.5500 29.0781 -1.3127 0.0446 0.3240  0.0290  0.0323 0.2342  0.0323 43.8699 
002370 20130930 5.0226 -200.3027 -737.7620 3.0019 -0.0209 0.2699 -0.0386 -0.0190 -0.0964 -0.0190 35.1696 
002393 20130930 2.3108 21.3287 16.8411 3.2078 0.1590 0.6495  0.3591  0.3295 0.1034  0.3295 66.9792 
002422 20130930 2.3011 18.7523 15.4236 1.7534 0.0726 0.4491  0.1920  0.1722 0.1322  0.1722 219.3120 
002437 20130930 4.9315 56.5184 52.4612 1.5637 0.0700 0.7674  0.1347  0.1422 1.0571  0.1422 113.4000 
002653 20130930 13.3723 46.9329 53.3025 2.3415 0.4212 0.7145  0.4425  0.4882 0.3671  0.4882 235.8229 
002693 20130930 3.1935 41.3297 38.7854 -7.6784 0.0693 0.6249  0.4072  0.4565 0.0800  0.4565 27.0720 
300006 20130930 5.2367 112.0564 60.3387 -16.2008 0.0406 0.3715  0.0542  0.0802 0.1406  0.0802 56.6032 
300016 20130930 5.0141 41.5183 35.1102 7.8817 0.1495 0.6945  0.2276  0.2076 0.1679  0.2076 27.1893 
300026 20130930 8.9274 46.4318 37.1856 1.0558 0.2022 0.8317  0.2072  0.1770 0.7922  0.1770 143.2314 
300086 20130930 1.6353 109.4084 56.0711 5.9406 0.0028 0.4658  0.1238  0.0917 -0.1995  0.0917 29.0000 
300110 20130930 2.9636 37.6105 22.9847 2.5999 0.0610 0.5250  0.1728  0.1443 0.5337  0.1443 40.2620 
300194 20130930 2.2659 96.6729 69.7392 -1.5238 0.0137 0.3165  0.1219  0.1087 -0.2244  0.1087 38.3085 
300199 20130930 8.6975 95.1120 82.4783 7.8161 0.0909 0.8133  0.4029  0.3731 0.2354  0.3731 92.6000 
300254 20130930 3.3275 69.2074 38.1987 3.6941 0.0562 0.6231  0.0952  0.0618 0.2605  0.0618 20.0700 
600062 20130930 2.3357 17.6355 11.1323 1.8080 0.1230 0.3112  0.1078  0.0908 0.0298  0.0908 116.5688 
600079 20130930 3.2838 35.7598 17.0296 -7.0877 0.0933 0.4182  0.1281  0.0723 0.2095  0.0723 139.2799 
600276 20130930 8.2268 41.4846 31.2832 2.7490 0.2097 0.7877  0.2512  0.2040 0.1232  0.2040 497.2969 
600297 20130930 2.1951 66.8464 49.0453 -1.9347 0.0411 0.2627  0.0722  0.0972 -0.5125  0.0972 20.3350 
600380 20130930 1.7856 30.1339 10.2166 0.7840 0.0768 0.6047  0.1231  0.0590 -0.0068  0.0590 72.6543 
600420 20130930 4.5755 33.8193 19.7438 2.0176 0.0709 0.3901  0.0803  0.0471 0.2715  0.0471 43.5053 
600513 20130930 4.1317 46.4486 32.1551 1.1443 0.0727 0.4439  0.0680  0.0608 0.9255  0.0608 17.0960 
600568 20130930 3.0480 59.0073 217.2803 -4.5461 0.0354 0.1229  0.0455  0.0717 0.3563  0.0717 31.0560 
600664 20130930 1.3296 22.9746 10.6580 -2.7641 0.0309 0.2795  0.0243  0.0192 0.0178  0.0192 109.2965 
600771 20130930 123.2274 13.6873 -134.3250 -0.1349 0.0067 0.5464 -0.1835  0.0054 -0.0652  0.0054 58.3189 
600789 20130930 1.9866 -32.9747 40.2872 0.7011 -0.0232 0.1285 -0.0554 -0.0307 -0.0761 -0.0307 29.1369 
600829 20130930 1.7848 55.0648 21.7546 -0.4271 -0.0692 0.3676 -0.0320 -0.0289 -0.2365 -0.0289 37.4608 

[33 rows x 11 columns] 
>>> df = ACEV(get_SW_code('sdht', sw_cls=3))[['PB','PE','EV_EBITDA','PEG','ROIC','mg_r','opr_pft_r','net_pft_r','sales_gr','net_pft_r','Ttl_mkv']].sort('EV_EBITDA') 
Traceback (most recent call last): 
    File "<stdin>", line 1, in <module> 
    File "D:\Python2.7\lib\site-packages\pandas\core\frame.py", line 2664, in sort 
    inplace=inplace) 
    File "D:\Python2.7\lib\site-packages\pandas\core\frame.py", line 2756, in sort_index 
    return self.take(indexer, axis=axis, convert=False, is_copy=False) 
    File "D:\Python2.7\lib\site-packages\pandas\core\generic.py", line 1126, in take 
    new_data = self._data.take(indices, axis=baxis) 
    File "D:\Python2.7\lib\site-packages\pandas\core\internals.py", line 3326, in take 
    ref_items=new_axes[0], axis=axis) 
    File "D:\Python2.7\lib\site-packages\pandas\core\internals.py", line 2304, in apply 
    do_integrity_check=do_integrity_check) 
    File "D:\Python2.7\lib\site-packages\pandas\core\internals.py", line 1957, in __init__ 
    self._set_ref_locs(do_refs=True) 
    File "D:\Python2.7\lib\site-packages\pandas\core\internals.py", line 2097, in _set_ref_locs 
    'have _ref_locs set' % (block, labels)) 
AssertionError: Cannot create BlockManager._ref_locs because block [FloatBlock: [PB], 1 x 33, dtype: float64] with duplicate items [Index([u'PB', u'PE', u'EV_EBITDA', u'PEG', u'ROIC', u'mg_r', u'opr_pft_r', u'net_pft_r', u'sales_gr', u'net_pft_r', u'Ttl_mkv'], dtype='object')] does not have _ref_locs set 
>>> 
>>> pd.__version__ 
'0.13.0-203-g4518630' 

只是想知道如何解決造成df.sort這個錯誤？

Answer 1

錯誤消息與所有BlockManager/_ref_locs東西有點混亂，但它似乎是因爲您選擇重複列（net_pft_r）：

df = ACEV(get_SW_code('sdht', sw_cls=3))[['PB','PE','EV_EBITDA','PEG','ROIC', 
'mg_r','opr_pft_r','net_pft_r','sales_gr','net_pft_r','Ttl_mkv']] 

例如：

>>> df = pd.DataFrame({"A": range(3), "B": range(3,6)}) 
>>> df[["A", "B", "A"]].sort("B") 
Traceback (most recent call last): 
    File "<ipython-input-53-12e1eca79c2b>", line 1, in <module> 
    df[["A", "B", "A"]].sort("B") 
    File "/usr/local/lib/python2.7/dist-packages/pandas-0.13.0_212_g65b9815-py2.7-linux-i686.egg/pandas/core/frame.py", line 2664, in sort 
    inplace=inplace) 
    File "/usr/local/lib/python2.7/dist-packages/pandas-0.13.0_212_g65b9815-py2.7-linux-i686.egg/pandas/core/frame.py", line 2756, in sort_index 
    return self.take(indexer, axis=axis, convert=False, is_copy=False) 
    File "/usr/local/lib/python2.7/dist-packages/pandas-0.13.0_212_g65b9815-py2.7-linux-i686.egg/pandas/core/generic.py", line 1126, in take 
    new_data = self._data.take(indices, axis=baxis) 
    File "/usr/local/lib/python2.7/dist-packages/pandas-0.13.0_212_g65b9815-py2.7-linux-i686.egg/pandas/core/internals.py", line 3326, in take 
    ref_items=new_axes[0], axis=axis) 
    File "/usr/local/lib/python2.7/dist-packages/pandas-0.13.0_212_g65b9815-py2.7-linux-i686.egg/pandas/core/internals.py", line 2304, in apply 
    do_integrity_check=do_integrity_check) 
    File "/usr/local/lib/python2.7/dist-packages/pandas-0.13.0_212_g65b9815-py2.7-linux-i686.egg/pandas/core/internals.py", line 1957, in __init__ 
    self._set_ref_locs(do_refs=True) 
    File "/usr/local/lib/python2.7/dist-packages/pandas-0.13.0_212_g65b9815-py2.7-linux-i686.egg/pandas/core/internals.py", line 2097, in _set_ref_locs 
    'have _ref_locs set' % (block, labels)) 
AssertionError: Cannot create BlockManager._ref_locs because block [IntBlock: [A], 1 x 3, dtype: int64] with duplicate items [Index([u'A', u'B', u'A'], dtype='object')] does not have _ref_locs set 

我不確定這是否是您的錯字，或者您是否有重複的列名或什麼，但如果您避免它，我認爲它應該起作用。