Method and system for maintaining the analytical quality of the micro-data, to ensure statistical disclosure restriction of categorical or continuous micro data (SDL). The new SDL method, in order to missing the (1) information, including the attribution, along with some of the adjustment, and to take (instead of the census) sample, the record is still in disclosure risk (2) with some adjustment for, instead of the original data set, utilize similarities between be disclosed subsets. Survey sampling is to reduce the financial cost in comparison to the census, with some of the loss of information. Similarly, if published a subset, but to reduce the disclosure cost in comparison to the whole database, with some of the loss of information. Therefore, the optimal survey sampling methods can be used for statistical disclosure sine. The method includes partitioning the database, and risk hierarchy, the optimum probability substitutions, optimal probability subsampling, and the optimal sampling weight calibration.