We have seen in non-RDBMS databases that you must know the relationship of the data in order to retrieve the data properly. These relationships are maintained by hidden pointers and references in the data. These pointers are created when the database is defined and loaded. If a field is added or a relationship changes, then the database must be reloaded and all programs using the records must be changed, even though the program does not use the new fields.RDBMS has none of these problems. You can add columns or use a new relationship at any time. Operators, such as JOIN, are used to compare actual values in the tables to establish relationships. You can tell your RDBMS what you want and the system will determine how to retrieve the data. This is called “data independence.”