System Level Design of Archive Manager which is to archive data previously maintained in rdms based tables. Give in the OOPS based design as well

I was recently asked in an interview to Design an archive manager which would archive data from two or more related rdms tables in conjucture to be available for being queried on various column fields at a future date.

I gave the solution of pushing view acheived by joining the tables at a common field into a nosql table like mongo as documents and sharding the db based on chronology of insertion.

Can you suggest better and efficient ways, as neither the interviewer nor I was satisfied by this implementation.