MapDB is an embedded Java database engine that provides concurrent Maps, Sets, and Queues backed by disk storage or off-heap memory.
MapDB addresses the need for persistent, high-performance data structures in Java applications by combining an embedded database engine with familiar Java collection interfaces. It allows developers to use Maps, Lists, and Queues that transparently persist to disk or allocate memory outside the Java heap, avoiding garbage collection pauses. The tool supports multiple deployment modes: as a drop-in replacement for standard collections, as an off-heap cache unaffected by garbage collection, as a multilevel cache with expiration and disk overflow capabilities, or as a lightweight relational database alternative with transactions, MVCC, and incremental backups.
MapDB suits applications that need to handle large datasets efficiently without the overhead of a separate database server. It works well for local data processing and filtering tasks where performance matters and garbage collection latency is a concern. The off-heap storage capability makes it particularly valuable for applications managing substantial in-memory collections that would otherwise trigger frequent garbage collection. Developers should choose this tool when they want collection-like semantics with persistence guarantees, or when they need to process huge quantities of data in reasonable time without deploying a full RDBMS.
The project maintains an extensive test suite with over a million test cases, though the default build runs only a tiny fraction to keep build time reasonable. The codebase is written in Kotlin and requires IntelliJ IDEA for development. MapDB comes with comprehensive documentation including a quick start guide and full API documentation available on its website.
