BUG: Formerly useful behavior now raises.

[danielballan]

I did this:

In [1]: df1 = DataFrame([1, 2, 3, 4, 5], index=[1, 2, 1, 2, 3])

In [2]: df2 = DataFrame([1, 2, 3], index=[1, 2, 3])

In [3]: df1.sub(df2)
Out[3]:
0
1 0
1 2
2 0
2 2
3 2

which is possible up to and including this commit. Then, this "bug fix" seems to have broken the above usage. It now raises like so:

ValueError                                Traceback (most recent call last)
<ipython-input-12-e7e549a4e818> in <module>()
----> 1 df1.sub(df2)

/Users/danielallan/Documents/Repos/pandas-danielballan/pandas/core/ops.pyc in f(self, other, axis, level, fill_value)
    710     def f(self, other, axis=default_axis, level=None, fill_value=None):
    711         if isinstance(other, pd.DataFrame):    # Another DataFrame
--> 712             return self._combine_frame(other, na_op, fill_value, level)
    713         elif isinstance(other, pd.Series):
    714             return self._combine_series(other, na_op, fill_value, axis, level)

/Users/danielallan/Documents/Repos/pandas-danielballan/pandas/core/frame.pyc in _combine_frame(self, other, func, fill_value, level)
   2759 
   2760     def _combine_frame(self, other, func, fill_value=None, level=None):
-> 2761         this, other = self.align(other, join='outer', level=level, copy=False)
   2762         new_index, new_columns = this.index, this.columns
   2763 

/Users/danielallan/Documents/Repos/pandas-danielballan/pandas/core/generic.pyc in align(self, other, join, axis, level, copy, fill_value, method, limit, fill_axis)
   2331                                      copy=copy, fill_value=fill_value,
   2332                                      method=method, limit=limit,
-> 2333                                      fill_axis=fill_axis)
   2334         elif isinstance(other, Series):
   2335             return self._align_series(other, join=join, axis=axis, level=level,

/Users/danielallan/Documents/Repos/pandas-danielballan/pandas/core/generic.pyc in _align_frame(self, other, join, axis, level, copy, fill_value, method, limit, fill_axis)
   2362         left = self._reindex_with_indexers({0: [join_index,   ilidx],
   2363                                             1: [join_columns, clidx]},
-> 2364                                            copy=copy, fill_value=fill_value)
   2365         right = other._reindex_with_indexers({0: [join_index,   iridx],
   2366                                               1: [join_columns, cridx]},

/Users/danielallan/Documents/Repos/pandas-danielballan/pandas/core/generic.pyc in _reindex_with_indexers(self, reindexers, method, fill_value, limit, copy, allow_dups)
   1320                 indexer = com._ensure_int64(indexer)
   1321                 new_data = new_data.reindex_indexer(index, indexer, axis=baxis,
-> 1322                                                     fill_value=fill_value, allow_dups=allow_dups)
   1323 
   1324             elif baxis == 0 and index is not None and index is not new_data.axes[baxis]:

/Users/danielallan/Documents/Repos/pandas-danielballan/pandas/core/internals.pyc in reindex_indexer(self, new_axis, indexer, axis, fill_value, allow_dups)
   2884         # trying to reindex on an axis with duplicates
   2885         if not allow_dups and not self.axes[axis].is_unique:
-> 2886             raise ValueError("cannot reindex from a duplicate axis")
   2887 
   2888         if axis == 0:

ValueError: cannot reindex from a duplicate axis

I found this very handy. Should it really raise? Can you suggest what I should be doing instead? Thanks.

[jreback]

that is disappointing! let me take a look

[jreback]

@danielballan oversight on that operation....see #5187
look right? (same behavior in 0.12)

[danielballan]

Yes, all fixed. Glad it has a test now. Thanks for handling this promptly.