Partial fancy indexing using ints does not work on MultiIndex

[wesm]

In [44]: import pandas.rpy.common as com

In [45]: cw = com.load_data('ChickWeight')

         Chick  weight
1  0.0   10.5   41.4  
   2.0   10.5   47.25 
   4.0   10.11  56.47 
   6.0   10.11  66.79 
   8.0   10.11  79.68 
   10.0  10.11  93.05 
   12.0  10.11  108.5 
   14.0  9.778  123.4 
   16.0  9.471  144.6 
   18.0  9.471  158.9 
   20.0  9.471  170.4 
   21.0  9.562  177.8 
2  0.0   25.5   40.7  
   2.0   25.5   49.4  
   4.0   25.5   59.8  
   6.0   25.5   75.4  
   8.0   25.5   91.7  

In [47]: cw.groupby(['Diet', 'Time']).mean().ix[1]
Out[47]: 
Chick     10.5
weight    47.25

[wesm]

fixed here (though in latest rpy the Diet column flows through as object dtype)

wesm@4e50981