_sanitize_column now reports proper duplicate error

[roblevy]

closes #7432
A very simple fix for GH7432. internals.py (line 3240) actually raises the correct exception (ValueError: cannot reindex from a duplicate axis) but the except: doesn't allow it to be raised. Here I've just allowed the exception itself to be raised since it's more useful to the user than the weird TypeError message.

[roblevy]

Here's what the situation in Issue 7432 now looks like:

In [4]: df['newcol'] = ser
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-4-acaa0ccb832e> in <module>()
----> 1 df['hello'] = ser

/home/rob/pandas_roblevy/pandas/core/frame.py in __setitem__(self, key, value)
   1853         else:
   1854             # set column
-> 1855             self._set_item(key, value)
   1856 
   1857     def _setitem_slice(self, key, value):

/home/rob/pandas_roblevy/pandas/core/frame.py in _set_item(self, key, value)
   1921         is_existing = key in self.columns
   1922         self._ensure_valid_index(value)
-> 1923         value = self._sanitize_column(key, value)
   1924         NDFrame._set_item(self, key, value)
   1925 

/home/rob/pandas_roblevy/pandas/core/frame.py in _sanitize_column(self, key, value)
   1964                     value = value.reindex(self.index).values
   1965                 except ValueError as e:
-> 1966                     raise e
   1967                 except:
   1968                     raise TypeError('incompatible index of inserted column '

ValueError: cannot reindex from a duplicate axis

[jreback]

yep..looks trivial...can you rebase (so that travis will then run).

[jreback]

thanks!

[jreback]

the importance of testing! turns out their are some other error messages here that are effectively raising as TypeError, fixed up here: 7ef4efa

[roblevy]

Amazing. Next time I do a pull request I'll add a test for whatever I add. This is a good learning experience for me...