The existing methods generalization and bucketization could not give a satisfactory result for privacy preserving on microdata. It results loss of information and some does not prevent the membership disclosure. In this paper, the author presents a new idea slicing, which partitions the data both horizontally and vertically. The author justify that slicing preserves the data integrity and gives the member protection even it can handle high dimensional data. This can be used in protection of attribute disclosure and develop an algorithm to obey the diversity requirement.