Getting started with Ecto (Elixir)
Working with Ecto.
- Purpose
- What’s Elixir?
- What’s Ecto?
- Creating the project
- Ecto tasks
- Mapping tables to elixir structures
- Testing the application
- The code
- Links of interest
Purpose
The purpose of this article is sharing with other geeks my experience developing an application with Elixir. A few months ago I just did some researches in coding with Erlang and to be honest I really enjoyed it writing the article First steps with Erlang, but this time I will go for a more real application which could provide me with some better feedback about this functional programming language.
What’s Elixir?
Elixir is a dynamic, functional language designed for building scalable and maintainable applications and run over the Erlang Virtual Machine
What’s Ecto?
Ecto is a database wrapper that makes the life easier for Elixir developers who need to make use of database in their applications. Ecto works with Postgresql, Mysql, MSSQL, SQLite3 and MongoDB.
In this tutotial we will be working with a Postgresql but we could do de same with any of the above databases.
Creating the project
By this point I assume you all have elixir installed on your computer. If not, you only need to got to the Elixir Siteand follow the installation steps.
Create the project
mix new -app <APP_NAME>
This will create some directories:
_build: This directory contains only compiled code and libraries.
config: Configuration for our application.
deps: Dependencies used by this application.
lib: The code of our project.
test: Unit testing for our application
mix.exs: In this file we define the dependencies and the application modules.
priv/repo: SQL scripts and Elixir migration scripts.
Adding dependencies
Since we are working with Postgresql we need to add the driver beside of the Ecto dependency.
So in the mix.exs file we add the below
...
def application do
[applications: [:logger, :postgrex, :ecto]]
end
defp deps do
[
{:postgrex, ">= 0.0.0"},
{:ecto, "~> 2.0.0"}
]
end
...
and to downlaod dependcy we run the command
mix deps.get
Configuring our database connection
First af all we create a Repo for our application. We create a new file in directory lib,
/lib/repo.exs
defmodule ElixirEctoTraining.Repo do
use Ecto.Repo,
otp_app: :elixir_ecto_training
end
and then we edit file /config/config.exs to define the database driver connection. For this tutorial we assume a postgresql instance run on port 3001 (we can make use of docker-compose.yml in docker directory)
config :elixir_ecto_training, ecto_repos: [ElixirEctoTraining.Repo]
config :elixir_ecto_training, ElixirEctoTraining.Repo,
adapter: Ecto.Adapters.Postgres,
database: "elixir_ecto_training_db",
username: "wesovi",
password: "wesoviPass",
hostname: "localhost",
port: "3001"
Ecto tasks
Ecto provides with samo taks which we can find running
mix ecto --help
Creating the database by running ecto.create
Defining the tables structure.
We need to create a file called /priv/lib/structure.sql. So far we will only create a table to store tracks.
CREATE SEQUENCE tracks_id_seq;
CREATE TABLE tracks (
id bigint NOT NULL DEFAULT nextval('tracks_id_seq'),
title VARCHAR,
inserted_at timestamp without time zone
);
ALTER SEQUENCE tracks_id_seq OWNED BY tracks.id;
ALTER TABLE ONLY tracks
ADD CONSTRAINT tracks PRIMARY KEY (id);
To create the tables we only need to run command ecto.load
We can check the table has been created by using any Postgresql Admin Tool,
Oooops… we forgot defining a column singer in table tracks… What can we do now?, We well generate a new migration for do this.
We need to edit new migration file in order to write our sql change.
defmodule ElixirEctoTraining.Repo.Migrations.AddColumnSingerToTableTracks do
use Ecto.Migration
def change do
alter table(:tracks) do
add :singer, :string
end
end
end
We can have a look at all the migrations and theis statuses by running mix ecto.migration
To make the change defined in the migration file we can do by mix ecto.migrate -v 20160717214057
And the new column singer has been added to the table tracks.
Mapping tables to elixir structures
We create a new file /lib/model.exs and define a new module as below:
defmodule ElixirEctoTraining.Track do
use Ecto.Schema
import Ecto.Query
schema "tracks" do
field :title
field :singer
field :inserted_at
end
end
Testing the application
We modify mix.exs file and add the below content
def application do
[
applications: [:logger, :postgrex, :ecto],
mod: {ElixirEctoTraining,[]}
]
end
In order to run our tests agains a empty database we will create an aliases in out mix.exs file
What’s an aliases Aliases allow us to perform some tasks that contain sub-tasks
...
def project do
[app: :elixir_ecto_training,
version: "0.1.0",
elixir: "~> 1.3",
build_embedded: Mix.env == :prod,
start_permanent: Mix.env == :prod,
deps: deps(),
aliases: aliases]
end
...
defp aliases do
[
"ecto.rebuild": [
"ecto.drop", "ecto.create", "ecto.load", "ecto.migrate", "test"
],
"test": [
"ecto.rebuild"
]
]
end
We can write a very basic test. The purpose of this test is check tha following: We have an empty table tracks and add a new record then we have 1 record.
So we create a file /test/elixir_ecto_training.ex and with the below content
defmodule ElixirEctoTrainingTest do
alias ElixirEctoTraining.Track
alias ElixirEctoTraining.Repo
import Ecto.Query
use ExUnit.Case, async: true
setup do
:ok = Ecto.Adapters.SQL.Sandbox.checkout(Repo)
Ecto.Adapters.SQL.Sandbox.mode(Repo, {:shared, self()})
end
test "track insertion" do
previousNumberOfTracks = List.first Repo.all(from u in Track, select: count(u.id))
Repo.insert! %Track{ title: "Runaway", singer: "John Bon Jovi" }
numberOfTracks = List.first Repo.all(from u in Track, select: count(u.id))
assert numberOfTracks == previousNumberOfTracks + 1
end
end
To run the tests we do
mix test
The code
As always the code can be found on wesovilabs repository
