Assignment to DataFrame through .ix with non-unique MultiIndex

[manuteleco]

Hi,

I think there might be a bug in the assignment to a DataFrame object through .ix when non-unique MultiIndexes are involved. The next sample code describes this issue:

In [1]: from pandas import DataFrame, MultiIndex

In [2]: 

In [2]: df = DataFrame([[1, 1, "x", "X"], [1, 1, "y", "Y"], [1, 2, "z", "Z"]],
   ...:                columns=list("ABCD"))

In [3]: df = df.set_index(list("AB"))

In [4]: df
Out[4]: 
     C  D
A B      
1 1  x  X
  1  y  Y
  2  z  Z

In [5]: 

In [5]: ix = MultiIndex.from_tuples([(1, 1)])

In [6]: df.ix[ix, "C"]  # Shows both values with index (1, 1)
Out[6]: 
A  B
1  1    x
   1    y
Name: C

In [7]: 

In [7]: df.ix[ix, "C"] = "_"  # Only sets one of the two values with index (1, 1): BUG?

In [8]: df
Out[8]: 
     C  D
A B      
1 1  x  X
  1  _  Y
  2  z  Z

Thanks and regards.
Manu.

[wesm]

thanks for the report. fixed this one