ORMlike

ORMlike is a software data access pattern that provides object-relational mapping–like capabilities without the breadth of a full object-relational mapper (ORM). It maps between in-memory domain objects and relational data sources, but generally relies on explicit SQL or a lightweight query API rather than fully automated query generation, identity tracking, and change propagation.

Typical features include a defined mapping between classes and tables, simple CRUD operations, type-safe query composition or direct SQL execution, and basic connection and transaction management. Many implementations emphasize predictable performance, transparency, and closer control over the generated queries, while avoiding some of the magical behavior associated with larger ORMs.

ORMlike tools are contrasted with full ORMs by offering less automation and fewer surprises, trading ease of use for performance and control. They are similar in spirit to micro-ORMs or data-mapper approaches, depending on the exact design. They also sit between raw SQL and a full ORM in terms of developer effort and abstraction level.

Benefits of ORMlike patterns include lower overhead and faster startup, easier debugging, compatibility with existing schemas, and more direct SQL optimization. Drawbacks include more boilerplate, less automatic relationship handling, reduced support for advanced features such as lazy loading and cascading, and potential for consistency gaps if the mapping layer is not carefully maintained.

Common contexts for ORMlike patterns include applications with performance constraints, complex legacy databases, or teams that prefer explicit SQL alongside object typing. Examples often described as micro-ORMs or SQL mappers include Dapper in .NET, JDBI or MyBatis-like approaches in Java, and Doobie-style SQL mapping in Scala. These tools typically encourage the repository pattern to encapsulate data access and keep domain models persistence-ignorant.