HADS
GroveDB is the first production-level implementation of a HADS, largely inspired by Database Outsourcing with Hierarchical Authenticated Data Structures.
HADS stands for "hierarchical authenticated data structure". Simply speaking, it's an ADS (authenticated data structure), such as a Merkle tree, which contains additional ADSs.
The HADS architecture is what enables GroveDB to be the first database to support proofs for complex queries. Trees can be created for secondary indexes, where each secondary index tree (such as for "city" in a data set of restaurants) has additional subtrees for each unique value (such as "New York City" and "Phoenix"), and the items in that subtree (all the restaurants in New York City or Phoenix) can refer to the records in the primary index tree, where the full data of each item (restaurant) is stored.
Note that the secondary index trees only contain references to items, rather than the items themselves, for storage space efficiency.
Benefits of a HADS:
Efficient proofs for complex queries
Efficient handling of duplicate values
Different ADS types can be used within the same HADS
Supports compound indexes