Set Up the Sample Application

How is Ecto organized?

Under the hood, Ecto is actually two separate packages; ecto and ecto_sql. The ecto package contains some of the core data manipulation features that are useful even if we’re not using a relational database. These include the Repo, Query, Schema, and Changeset modules, among others.

The ecto_sql package, on the other hand, contains modules specifically needed to communicate with relational databases. These include the various database-specific adapters, migrations, and so forth.

Ecto with relational database

The ecto_sql package includes ecto as a dependency. If we’re using Ecto to work with a relational database, we need to include ecto_sql in our dependencies to get ecto in the process. Suppose we’re not working with a relational database and want to take advantage of some of the data manipulation features that Ecto offers, such as validations, though. In that case, we can include ecto rather than ecto_sql to make our dependency tree a little lighter.