Subsetting creates smaller, referentially-intact copies of large databases.

By default, subsetting creates a subset that is 10% of the source database (up to a maximum of 1GB).

You can control what data is included in two ways: by specifying a desired size, or by choosing a starting table.

Desired Size

Specify the desired size of your subset in disk space (e.g., 5GB). The subsetter uses statistical sampling to select representative data across all tables to achieve approximately the requested size.

This approach gives you easy control over infrastructure resources, but less control over exactly which data is included.

See desiredSize in the subsetting options file reference.

Starting Table

Specify a table and optional filter clause to start from. The subsetter follows foreign key relationships from this starting point, collecting all related data from dependent tables.

This approach gives you precise control over which data is included, but requires knowledge of your database structure.

See startingTables in the subsetting options file reference.

Advanced Customization

For advanced use-cases, you can use an options file to specify multiple starting tables, static data tables, and manual relationships.

More information is detailed on our subsetting options file page.