In this paper, the authors present two approaches one of that provided service for every incoming tuple for insertion even if that tuple fails the test of secure protocol of the private updating anonymous database by periodically extracting k-anonymous part of pending tuple set (i.e. all tuples that fail insertion). If k-anonymous part not available in pending tuple set means all tuples in pending tuple are unique then by using another approach to populate the pending tuple set as an anonymous by using the k-mean clustering algorithm and make every tuple identical in one cluster by suppressing as few as possible.