Should GroupBy.transform exclude nuisance columns?

[wesm]

In [46]: print df; df.groupby('key1').transform(lambda x: x.mean())
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
/home/wesm/Dropbox/book/svn/book_scripts/notebooks/<ipython-input-46-656cf72f4323> in <module>()
      1 print df
----> 2 df.groupby('key1').transform(lambda x: x.mean())

/home/wesm/code/pandas/pandas/core/groupby.pyc in transform(self, func, *args, **kwargs)
   1639                     group.T.values[:] = res
   1640                 else:
-> 1641                     group.values[:] = res
   1642 
   1643                 applied.append(group)

ValueError: operands could not be broadcast together with shapes (3,4) (2) 

      data1     data2 key1 key2
0  0.984916  0.904550    a  one
1 -1.879638  0.585887    a  two
2  0.457879 -0.721776    b  one
3  1.585887  0.050336    b  two
4 -0.701547  0.525808    a  one