Allow concat names arg to rename existing MultiIndex level.

[lodagro]

In [67]: a = pandas.DataFrame(np.random.rand(3,3),
   ....:                      columns=list('ABC'),
   ....:                      index=pandas.Index(list('abc'), name='index_a'))

In [68]: b = pandas.DataFrame(np.random.rand(3,3),
   ....:                      columns=list('ABC'),
   ....:                      index=pandas.Index(list('abc'), name='index_b'))

In [69]: a
Out[69]:
                A         B         C
index_a
a        0.734551  0.364354  0.989970
b        0.356027  0.625983  0.855296
c        0.895759  0.144406  0.068718

In [70]: b
Out[70]:
                A         B         C
index_b
a        0.130305  0.981312  0.541476
b        0.046006  0.369983  0.822214
c        0.621787  0.047847  0.420459

In [71]: c = pandas.concat([a, b], keys=['key0', 'key1'], names=['lvl0'])

In [72]: c
Out[72]:
                     A         B         C
lvl0 index_a
key0 a        0.734551  0.364354  0.989970
     b        0.356027  0.625983  0.855296
     c        0.895759  0.144406  0.068718
key1 a        0.130305  0.981312  0.541476
     b        0.046006  0.369983  0.822214
     c        0.621787  0.047847  0.420459

In [73]: c = pandas.concat([a, b], keys=['key0', 'key1'], names=['lvl0', 'lvl1'])
---------------------------------------------------------------------------
AssertionError                            Traceback (most recent call last)
...
AssertionError:

[wesm]

This is done in git master