Introduction: With many countries introducing rules to safeguard the personal data stored in the database, it is diffucult for the developer to replicate a user database in its office. With this program you can have the same client database, but without the private data. Each database table will be saved in a bunch of dataset files to avoid occuping excessively the memory during the export phase. Each exported table has the following attributes: set of key fields, and for each field optionally an operator and ids to eventually limit the output data. This is a filter. es: where id < xxx and id > yyyy for each field in the table we can apply a rule: 1- Compose the value in a set of predefined values with the same initial (to maintain the same distribution as the original tabel) es: prename: M->mario,michele,massimo, and a counter. If there no an appropiathe choice, then program can 'invent' a value 2- Picture with random or predefined data (es: ssn) 3- A special rule ( es: codice fiscale italiano) with a checksum ,etc. For this maybe a plugin interface has to be declared 4- A casual date +- a delta from the original one ( birth date ) 5- A composition (address: via,street, largo ,a set of names , and a number ) 6- A code from another table ( es: city table ). A special treatment apply to the random fields which are unique in its table or foreign keys of another table. Program invocation arguments ---------------------------- From command line: Commands (mandatory): -e: export operation (not yet done) -i: import operation (not yet done) Options: -c file: configuration file (not yet done) -n file: connection file (not yet done) -o directory: output directory (not yet done) -z: compress the files (not yet done) Misc: -h: help (not yet done) -v: version (not yet done)