inserting new column with DataFrame.ix buggy

[changhiskhan]

In [22]: df
Out[22]: 
          A         A         B         C
0  0.148734 -0.041765  1.575866  1.422524
1  0.188963  2.086219  1.595425  0.539364
2  1.498709 -0.338857 -0.210994  0.984697
3 -0.197789 -0.093690 -0.421010  0.710587

In [23]: df.ix[:, 'D'] = ['X', 'Y', 'Z', 'V']
---------------------------------------------------------------------------
KeyError                                  Traceback (most recent call last)
<ipython-input-23-e2003f3ca30c> in <module>()
----> 1 df.ix[:, 'D'] = ['X', 'Y', 'Z', 'V']

/Users/changshe/code/pandas/pandas/core/indexing.pyc in __setitem__(self, key, value)
     69                 raise IndexingError('only tuples of length <= %d supported',
     70                                     self.ndim)
---> 71             indexer = self._convert_tuple(key)
     72         else:
     73             indexer = self._convert_to_indexer(key)

/Users/changshe/code/pandas/pandas/core/indexing.pyc in _convert_tuple(self, key)
     78         keyidx = []
     79         for i, k in enumerate(key):
---> 80             idx = self._convert_to_indexer(k, axis=i)
     81             keyidx.append(idx)
     82         return tuple(keyidx)

/Users/changshe/code/pandas/pandas/core/indexing.pyc in _convert_to_indexer(self, obj, axis)
    535                 return indexer
    536         else:
--> 537             return labels.get_loc(obj)
    538 
    539     def _tuplify(self, loc):

/Users/changshe/code/pandas/pandas/core/index.pyc in get_loc(self, key)
    716         loc : int if unique index, possibly slice or mask if not
    717         """
--> 718         return self._engine.get_loc(key)
    719 
    720     def get_value(self, series, key):

/Users/changshe/code/pandas/pandas/index.so in pandas.index.IndexEngine.get_loc (pandas/index.c:3222)()

/Users/changshe/code/pandas/pandas/index.so in pandas.index.IndexEngine.get_loc (pandas/index.c:3053)()

/Users/changshe/code/pandas/pandas/index.so in pandas.index.IndexEngine._get_loc_duplicates (pandas/index.c:3389)()

KeyError: 'D'

[changhiskhan]

This works fine:

In [24]: df['D'] = ['X', 'Y', 'Z', 'V']

In [25]: df
Out[25]: 
          A         A         B         C  D
0  0.148734 -0.041765  1.575866  1.422524  X
1  0.188963  2.086219  1.595425  0.539364  Y
2  1.498709 -0.338857 -0.210994  0.984697  Z
3 -0.197789 -0.093690 -0.421010  0.710587  V

[jreback]

this looks similar to #2578 FYI

[wesm]

FWIW I don't believe this was ever a supported/intended API

[jreback]

I actually like this. it's very consistent and I thinks concise way of getting / setting (sort of a more advanced way but that's always good). and in higher dimensions, the only 'natural' way of doing it.

[jreback]

pushing to 0.12

[jreback]

closing as a dup of #2578 which is implemented