We can map SQL columns to class definitions so that we can create proper Python object instances from data in a database. SQLite includes a Binary Large Object (BLOB) data type. We can pickle our Python objects and store them in the BLOB columns. We can work out a string representation of our Python objects (for example, using the JSON or YAML notation) and use SQLite text columns too.
This technique must be used cautiously because it effectively defeats SQL processing. A BLOB column cannot be used for SQL DML operations. We can't index it or use it in the search criteria of DML statements.
SQLite BLOB mapping should be reserved for objects where it's acceptable to be opaque to the surrounding SQL processing. The most common examples are media objects such as videos, still images, or sound clips. SQL is biased towards text...
