"Key" column pandas implementation of missingness

[pwochner]

At the moment, the pandas version treats the Key column as a regular column and includes it in the missingness computation. The postgres version doesn't. The name of the key column could be provided as optional input argument and this column could be used as index in Dataframe.