Essential SQLAlchemy

Chapter 6. Building an Object Mapper

Atop the SQL expression language, SQLAlchemy provides an object-relational mapper (ORM). The purpose of an ORM is to provide a convenient way to store your application data objects in a relational database. Generally, an ORM will provide a way to define the method of storing your object in the database. This chapter focuses on the SQLAlchemy methods that do this.

Introduction to ORMs

ORMs provide methods of updating the database by using your application objects. For instance, to update a column in a mapped table in SQLAlchemy, you merely have to update the object, and SQLAlchemy will take care of making sure that the change is reflected in the database. ORMs also allow you to construct application objects based on database queries. Chapter 7 will focus on how to use SQLAlchemy’s ORM to update and query objects in the database.

Design Concepts in the ORM

There are two major patterns used in the ORM you should become familiar with in order to understand how to best use the ORM. These are the data mapper pattern and the unit of work pattern.

The data mapper pattern

In the data mapper pattern (shown in Figure 6-1), database tables, views, and other “selectable” objects are mapped onto “plain old Python objects” (POPOs) by “mapper” objects. This is different from the “active record” pattern (shown in Figure 6-2), where the objects themselves are responsible for mapping themselves to database views. The data mapper pattern can, of course, be used to emulate ...