J2EE Design Patterns

Data Access Patterns

The most common J2EE persistence mechanism is a relational database management system, or RDBMS. Therefore, we focus primarily on patterns that apply to relational databases.

Java, both in and out of J2EE, is chock full of database technologies. Enterprise JavaBeans were designed with the database backend in mind, and the JDBC API, which is part of the J2SE package but heavily used in J2EE environments, provides a powerful, low-level approach for connecting any Java object to an external database. Other technologies, such as Java Data Objects, provide a higher-level abstraction of the persistent data, handling much of the mapping between database tables and Java objects.

Of course, databases are not the only persistence mechanism available. Simpler applications can persist data to files on disk in text or binary formats, or to XML. Java objects themselves can be serialized directly to disk and reloaded later, although this approach makes it difficult for human beings to access and manipulate the data independently from the program that created it. Collectively, these persistence mechanisms are referred to as data sources .

Whatever data source we use, there are problems when we embed persistence logic directly into business components. When the business object is responsible for persistence, it means we’ve given two responsibilities to a single class. Doing so isn’t intrinsically evil, but it can be problematic in several ways. First, the full persistence layer ...