When seeding a sandbox, you often need a deliberate mix of records. For example, you might want to pull Accounts from two different cities within a single deployment, applying unique constraints to each group. Filter subsets let you build that exact mix in one go. Because each subset is filtered independently, you can layer extra criteria or set specific record limits for one city without affecting the other.
How to define filter subsets
To open the filter window, click the filter icon on an object selected for deployment.
From here, you'll be able to define data subsets to filter by using the '+ Define subsets'-option.
Any top-level filters, such as the one on AnnualRevenue in the example above, will apply to all subsets defined for the object. For each subset, you can set a record limit and add one or more filters to define which records it includes.
As seen in the demonstration above, this would deploy:
Up to 50 records that have an annual revenue of over 100,000, and London as their billing city.
Up to 50 records that have an annual revenue of over 100,000, and New York as their billing city.
The matching record count is shown alongside each subset, so you can see how many records match your filters before deploying.