You can do it with the row_number() function. Partition the data according to status column and order it by state column in descending order. Then you can take only the columns that state is null and row_number() function returned 1 or the rows that state column did not have null as a value in it regardless of the value of row_number() function.

Adi