Essential SQLAlchemy

Chapter 4. SQLAlchemy Type Engines

This chapter introduces the SQLAlchemy type system. It covers the built-in types provided by SQLAlchemy: database-independent types and database-specific types. It then tells you how to create your own custom types for use in mapping application data onto your database schema.

Type System Overview

When defining the MetaData used by your application, it is necessary to supply the SQL data type used by each column of each table (unless the tables are defined with autoload=True, in which case SQLAlchemy provides the data types for you). These SQL data types are actually instances of SQLAlchemy-provided classes known as TypeEngines. TypeEngine objects convert Python values to native database values and vice versa. For instance, String(40) is an instance of a TypeEngine that represents a VARCHAR(40). TypeEngines also supply SQL text for use when creating tables using metadata.create_all() or table.create().

SQLAlchemy provides three different ways of constructing types for use in your application. First, it provides a set of generic TypeEngines, which are fairly portable across different database engines. Second, it provides database server-specific TypeEngines, which can be used to exploit particular types supported by certain databases. Third, SQLAlchemy allows you to define application-specific custom TypeEngines if you wish to further customize object conversion to/from the database.

Built-in Types

SQLAlchemy provides a fairly complete set of built-in ...