Ecto Transactions

Redactado por: Felipe Juarez

It has been a couple of months since I wrote something and was about phoenix and bootstrap. Today I’m going to talk about elixir, ecto and how to manage transactions

Well in this post we are going to skip ecto setup (if for some way, it was need it, please leave a comment and I will make a post for it, but I think his documentation makes a fairly good example for it)

With that in mind, these are the tables that we are going to use.

  • Project table

    defmodule EctoTransactions.Project do
      use Ecto.Schema
      import Ecto.Changeset
      schema "projects" do
        field :name
        has_many :tasks, EctoTransactions.Task
      def changeset(struct, params \\ %{}) do
        |> cast(params, [:name])
        |> validate_required([:name])

  • Task table

    defmodule EctoTransactions.Task do
      use Ecto.Schema
      import Ecto.Changeset
      schema "tasks" do
        field :description
        belongs_to :project, EctoTransactions.Project
      def changeset(struct, params \\ %{}) do
        |> cast(params, [:description, :project_id])
        |> validate_required([:description, :project_id])

With both tables we can see that Task depends from Project. So we are going to create a project with a task and we are going to see how can we handle a creation like that.

The first thing we should know is, how can we create a transaction? With this instruction Repo.transaction(fn -> end) Between arrow and end we put the code that we need. And for making a rollback we use Repo.rollback(value) and between parenthesis we can indicate the reason of why we are making that decision.

So we are going to create a module for handling this

  defmodule EctoTransactions.App do

    import Ecto
    alias EctoTransactions.Repo
    alias EctoTransactions.Project
    alias EctoTransactions.Task

    # creating project and task
    def create_project_with_task(project_params, task_params) do
      Repo.transaction(fn ->
        |> create_project_with_params
        |> add_task_to_project(task_params)

    # creating project
    defp create_project_with_params(project_params) do
      |> Project.changeset(project_params)
      |> Repo.insert

    # adding a task to a valid project
    defp add_task_to_project({:ok, project}, task_params) do
      changeset = project
                  |> build_assoc(:tasks)
                  |> Task.changeset(task_params)

      case Repo.insert(changeset) do
        {:ok, _task} -> project
        {:error, changeset} -> Repo.rollback(:task) # Rollback transaction for invalid task data
    defp add_task_to_project({:error, _changeset}, _task_params),
    do: Repo.rollback(:project) # Rollback transaction for invalid project data


So if we use iex -S mix and passing some args to this we can obtain the following results

  • Passing good options
    iex(1)> project_params = %{name: "Explain ecto transactions"}
    %{name: "Explain ecto transactions"}
    iex(2)> task_params = %{description: "Should create task for example"}
    %{description: "Should create task for example"}
    iex(3)> EctoTransactions.App.create_project_with_task(project_params, task_params)
    22:28:59.905 [debug] QUERY OK db=0.2ms queue=0.1ms
    begin []
    22:28:59.917 [debug] QUERY OK db=0.9ms
    INSERT INTO "projects" ("name","inserted_at","updated_at") VALUES ($1,$2,$3) RETURNING "id" ["Explain ecto transactions", {{2016, 9, 21}, {3, 28, 59, 0}}, {{2016, 9, 21}, {3, 28, 59, 0}}]
    22:28:59.921 [debug] QUERY OK db=3.1ms
    INSERT INTO "tasks" ("description","project_id","inserted_at","updated_at") VALUES ($1,$2,$3,$4) RETURNING "id" ["Should create task for example", 1, {{2016, 9, 21}, {3, 28, 59, 0}}, {{2016, 9, 21}, {3, 28, 59, 0}}]
    22:28:59.922 [debug] QUERY OK db=0.3ms
    commit []
     %EctoTransactions.Project{__meta__: #Ecto.Schema.Metadata<:loaded, "projects">,
      id: 1, inserted_at: #Ecto.DateTime<2016-09-21 03:28:59>,
      name: "Explain ecto transactions",
      tasks: #Ecto.Association.NotLoaded<association :tasks is not loaded>,
      updated_at: #Ecto.DateTime<2016-09-21 03:28:59>}}

ecto_simple> select * from projects; select * from tasks;
|   id | name                      | inserted_at         | updated_at          |
|    1 | Explain ecto transactions | 2016-09-21 03:28:59 | 2016-09-21 03:28:59 |
|   id | description                    |   project_id | inserted_at         | updated_at          |
|    1 | Should create task for example |            1 | 2016-09-21 03:28:59 | 2016-09-21 03:28:59 |
Time: 0.003s

And for bad data we can obtain next results:

iex(4)> project_params = %{name: 1}
%{name: 1}
iex(5)> task_params = %{description: nil}
%{description: nil}
iex(6)> EctoTransactions.App.create_project_with_task(project_params, task_params)

22:49:25.863 [debug] QUERY OK db=0.2ms
begin []

22:49:25.865 [debug] QUERY OK db=0.2ms
rollback []
{:error, :project}

iex(7)> project_params = %{name: "Explain ecto transactions"}
%{name: "Explain ecto transactions"}
iex(8)> EctoTransactions.App.create_project_with_task(project_params, task_params)

22:49:53.345 [debug] QUERY OK db=0.3ms queue=0.1ms
begin []

22:49:53.346 [debug] QUERY OK db=1.2ms
INSERT INTO "projects" ("name","inserted_at","updated_at") VALUES ($1,$2,$3) RETURNING "id" ["Explain ecto transactions", {{2016, 9, 21}, {3, 49, 53, 0}}, {{2016, 9, 21}, {3, 49, 53, 0}}]
#Ecto.Changeset<action: :insert, changes: %{},
 errors: [description: {"can't be blank", []}], data: #EctoTransactions.Task<>,
 valid?: false>

22:49:53.348 [debug] QUERY OK db=0.2ms
rollback []
{:error, :task}

If you read with attention you will see two tupples {:error, :project} and {:error, :task} with this you can make the assumptions that you want, of course you could say a lot more and instead of returning tupples you return other things with more sense for example change :project or :task for his respective changeset and you should see other things.

By the way when use Repo.transaction you should return the implicit data (that is the schema you are saving) that’s because Repo.transaction return a tupple consisting of an :ok atom and the respective last instruction. And if you use Repo.insert as returning data you will see a tupple of tupple {:ok, {:ok, data}} or {:ok, {:error, data}} (thats because the return of an insert is {:ok, data} or {:error, data}) and that’s why we use Repo.rollback for forcing a tupple consisting of {:error, reason}

That’s all folks! I hope this can help you. And remember Good Luck, Have Fun! and GG!

comments powered by Disqus