Pandas - calculate a rolling median on a dataframe that has a non-unique date index

Question

I have the following dataframe, which has a non-unique index using dates:

                column
2019-01-07         NaN    
2019-01-08         NaN
2019-01-08