# Database communication in Ecto
If you created your project using mix phx.new without using the --no-ecto flag, your business logic application shoud contain a repo in lib/project/repo.ex. This repo is your entrypoint for communication between your app and database and it's running alongside your business logic application as a separate process. The machinery behind then takes care of things like database connection pooling.
# Configuration
The database connection configuration is stored in the config/dev.exs file for development and in the config/prod.secret.exs for production under the Project.Repo section. You can either configure username, password and host separately or configure all at once by setting the url key to connection string.
The pool_size option configures number of open connections held in the connection pool managed by the repo.
# Creating table schema
To be able to execute database actions using your repo, you need to create a schema for each table in your database. You can create the schema using mix phx.gen.schema command if you're using the code-first database generation approach. This command will create your database schema from information provided and will also create a migration script that will alter your database so that its structure mimics the generated schema.
Since this guide is going to use the database-first approach, we will model our database separately so that we can use our database engine's potential to the fullest and then we will write our table schemas by hand.
Down below is an example of our handwritten RentToday.Film schema module. Notice a few important details. First of all, we didn't define the primary key inside the schema, but rather outside using the @primary_key {:field_name, :field_type, options} attribute. This is necessary if your primary key field name is other than id or the type is different than :id (which is integer in Elixir). This all is necessary because by default Ecto handles everything around primary key for you and you don't need to include it in the schema.
The second attribute used here, @schema is not about our Ecto schema, but rather database schema we are going to use. This is especially important if you are using Postgresql database - here you define the schema where your table is located.
defmodule RentToday.Film do
use Ecto.Schema
import Ecto.Changeset
@primary_key {:film_id, :integer, []}
@schema "public"
schema "film" do
field :title, :string
field :description, :string
field :release_year, :integer
field :created, :utc_datetime
field :created_by, :string
end
def changeset(film, attrs) do
film
|> cast(attrs, [:title, :description, :release_year, :created, :created_by])
|> validate_required([:title, :description, :release_year, :created, :created_by])
end
end