While evaluating access requests, Ranger policy engine picks policies based on the resource value specified in the access request. Currently access-resource abstraction only supports a single value for each resource-type - like database/table/column. Authorization of access to some resources might require the policy engine to pick policies based on multiple values for a resource.
For example, consider access authorization for an entity in Apache Atlas. An entity has a specific-type and a number of super-types - example: type=database super-types=[dataset, asset]. While authorizing access to a database entity, policies specified for its super-types, dataset and asset, should also be evaluated.
To enable such usecases, Ranger policy evaluation needs to be enhanced to support a list of value for a resource. Policies that match for any of the given values should be evaluated to determine the access result. Note that this enhancement doesn't require any updates to the policy model; the changes are needed only in the policy-engine.