All these questions were straightforward, HashTable is threadsafe but it is obsolete while HashMap implements Map interface which supports maintainability and reusability. For index we used to enhance the selection performance but we should add index to columns that doesn't have many insertion or deletion because these operations needs to update both the table and the indices. PK and FK are used for integrity, so we won't have inconsistency in our data and whenever we found a partial functional dependency we split the relation into two relations where the determinant will be the PK in the first relation and the dependent will be the FK in the second relation.