API: groupby-apply with on a DataFrame returning a Series can have misleading inference

[jreback]

xref #3599. Should prob not do this kind of inference.

In [83]: df1 = DataFrame([{"val1": 1, "val2" : 20}, {"val1":1, "val2": 19}, 
    ...:                  {"val1":2, "val2": 27}, {"val1":2, "val2": 12}])
    ...:                  

In [84]: df2 = DataFrame([{"val1": 1, "val2" : 20}, {"val1":1, "val2": 19}, 
    ...:                  {"val1":1, "val2": 27}, {"val1":1, "val2": 12}])
    ...:                  

In [85]: df1.groupby('val1').apply(lambda x: x.val2-x.val2.mean())
Out[85]: 
val1   
1     0    0.5
      1   -0.5
2     2    7.5
      3   -7.5
Name: val2, dtype: float64

In [86]: df2.groupby('val1').apply(lambda x: x.val2-x.val2.mean())
Out[86]: 
val2    0    1    2    3
val1                    
1     0.5 -0.5  7.5 -7.5

In [87]: df2.groupby('val1', squeeze=True).apply(lambda x: x.val2-x.val2.mean())
Out[87]: 
0    0.5
1   -0.5
2    7.5
3   -7.5
Name: 1, dtype: float64

[87] should a) have the correct index
but this should just work w/o the squeeze kwarg