A NoSQL database stores data in formats other than relational tables. Its flexible schema allows data to be persisted without fixing the complete schema in advance, supporting applications with rapidly changing requirements. section 18.1
Common NoSQL models include:
- Document-oriented databases store JSON-like documents; MongoDB organizes them into databases, collections, and documents. section 17.1 section 18.1
- Graph-oriented databases store nodes and edges, as in Neo4j. section 18.1
- Key/value databases use map-like keys and values, as in Redis; wide-column stores permit varying column names and formats, as in Apache Cassandra. section 18.1
Persistence frameworks can extend application portability from relational databases to NoSQL systems. Hibernate OGM reuses Hibernate Core, JPA, and JPQL so applications can target MongoDB or Neo4j with configuration changes, although portability can affect execution speed. section 18.0 section 18.1
JPA access to MongoDB
A JPA application can initialize an EntityManagerFactory, begin a transaction, use entityManager.find(User.class, id), and query with JPQL. Hibernate OGM maps these operations to MongoDB, where the persisted entities appear as collections and documents. section 18.2.3
Table 17.1 Compared terminology: RDBMS vs. MongoDB (view table figure)
Compare relational terminology with MongoDB's database, collection, document, and field model. section 17.1
