Datadmapper

Datadmapper is a software concept and, in some ecosystems, a library that implements the Data Mapper design pattern. It provides a persistence-ignorant domain model while a separate set of mapper objects handles translation between in-memory objects and a persistent store, such as a relational database.

In a datadmapper-based architecture, domain models contain business logic but no persistence concerns. Mappers or repositories load and save these objects by reading from and writing to the database, translating between domain objects and database rows or documents. A Unit of Work often coordinates multi-step updates to ensure consistency across related changes.

Key features include identity mapping (one domain object per database identity), change tracking, and often support for lazy loading, transactions, and mapping configurations that describe how properties map to columns and how associations are represented. Mappers may be configured declaratively (annotations, XML, or YAML) or procedurally, and can operate with raw SQL, a query builder, or an object-relational mapper that follows the Data Mapper pattern. The approach promotes separation of concerns and testability by keeping domain logic persistence-agnostic.

Datadmapper is typically chosen when domain models should remain independent of storage details, when multiple data sources are used, or when testability and clean separation are priorities. It introduces more boilerplate and learning curve compared with Active Record-style approaches.

The Data Mapper concept was popularized by Martin Fowler and remains a core pattern in enterprise architectures. Some frameworks implement it under the name datadmapper or as a module within larger data access layers.