I am trying to do a data pump but once it is finished the data is almost useless because I don't need 1,000 records per-table, but as many as meaningful are, the children records derived from smaller amounts of records in its corresponding Mater or Parent tables

So If have for instance ... let's say Businesses (1) and Stores (2) so I'd like to pump data just for these entities not 1,000 business and 1,000 stores but 1,000 billing records, 10,000 stock count records, 8,000 items catalog for each store ... derived from previous node ... like a tree structure of data-pumps (data-pump01, data-pump02, data-pump03, ... , data-pump99 ... and so on

It would be good if we could go increasingly seeding data to the master tables and later going on adding non-initially-added-tables so the newest data-pump will pump data into the newly selected tables using the values alredy stored in the database via the initial data-pump and in that way using the data stored in those previous records to create the children-ones

Just an idea !!!

P.S.: FOREIGN LANGUAGE DICTIONARIES WILL BE GREAT TOO (Caps just for distinguishing the 2 different ideas written here)