[DF] Update case when null handling for integers

[ayushdg]

Is your feature request related to a problem? Please describe. The return types for case when without else should match the data type of the variable in the then condition. In cases where then is an integer, the else should be a pandas nullable type rather than nan (which casts the whole column to float)

Describe the solution you'd like

Describe alternatives you've considered

Additional context