VIG provides a command line parameter that restricts the data to be generated to a granularity up to the level of a single column in the database.
--tables <string> (default: ) Restrict the generation to a list of tables. E.g., --tables=table1,table2,table3,etc.
--columns <string> (default: ) Restrict the generation to a list of columns. E.g., --columns=table1.col1,table2.col2,etc.
Assume you have 10 machines available, and you want to generate 10 tables (table1, …, table10). Then you can install a copy of VIG in each machine, and launch for the machine n the following command:
java -jar vig-version.jar --tables=tablen
The above command will force the machine n to only produce data for the table tablen. The produced CSV files can then be imported in the target RDBMS in the usual way.
The procedure with the parameter --columns is the same, with the only difference that the CSV files generated for each column must then be merged into a single CSV file that respects the order of the columns in the table. In UNIX systems, this is usually done through the paste command. Future releases of VIG will provide a command line utility that achieves the same result in an automatized way.