Issue Seeding With Ecto

I have a Phoenix app which is using ecto/postgres. I have setup two schemas (User has many Groups) and a seed file (contents below). I am running into an error when I add the group (which is supposed to have a user). I feel like the user association, where the constraint goes (requiring user_id vs assoc_constraint(:user)) and then how the seeding flows I am confused on.

Error I am getting

** (Ecto.InvalidChangesetError) could not perform insert because changeset is invalid.


    %{user_id: [{"can't be blank", [validation: :required]}]}

Applied changes

    %{description: "Random text", name: "Family"}


      "description" => "Random text",
      "name" => "Family",
      "user" => %XXXX.Accounts.User{
        __meta__: #Ecto.Schema.Metadata<:loaded, "users">,
        email: "",
        groups: #Ecto.Association.NotLoaded<association :groups is not loaded>,
        id: 1,
        inserted_at: ~N[2020-05-19 22:43:11],
        name: "Example User",
        password: "secret",
        password_hash: "XXXXXXXX",
        phonenumber: "1 203-247-7113",
        updated_at: ~N[2020-05-19 22:43:11],
        username: "exampleuser"


      action: :insert,
      changes: %{description: "Random text", name: "Family"},
      errors: [user_id: {"can't be blank", [validation: :required]}],
      data: #XXXX.People.Group<>,
      valid?: false

User Schema

defmodule XXXX.Accounts.User do
  use Ecto.Schema
  import Ecto.Changeset

  schema "users" do
    field :email, :string
    field :name, :string
    field :password_hash, :string
    field :password, :string, virtual: true
    field :phonenumber, :string
    field :username, :string

    has_many :groups, XXXX.People.Group


  def changeset(user, attrs) do
    |> cast(attrs, [:name, :username, :email, :password, :phonenumber])
    |> validate_required([:name, :username, :email, :password, :phonenumber])
    |> validate_length(:username, min: 4)
    |> validate_length(:password, min: 4)
    |> unique_constraint(:username)
    |> unique_constraint(:email)
    |> unique_constraint(:phonenumber)
    |> hash_password()

  defp hash_password(changeset) do
    case changeset do
      %Ecto.Changeset{valid?: true, changes: %{password: password}} ->
        put_change(changeset, :password_hash, Pbkdf2.hash_pwd_salt(password))
      _ ->

User Migration

defmodule XXXX.Repo.Migrations.CreateUsers do
  use Ecto.Migration

  def change do
    create table(:users) do
      add :name, :string, null: false
      add :username, :string, null: false
      add :email, :string, null: false
      add :password_hash, :string, null: false
      add :phonenumber, :string, null: false


    create unique_index(:users, [:username, :email, :phonenumber])

Group Schema

defmodule XXXX.People.Group do
  use Ecto.Schema
  import Ecto.Changeset

  schema "groups" do
    field :description, :string
    field :name, :string

    belongs_to :user, XXXX.Accounts.User


  def changeset(group, attrs) do
    required_fields = [:name, :user_id]
    optional_fields = [:description]

    |> cast(attrs, required_fields ++ optional_fields)
    |> validate_required(required_fields)
    |> assoc_constraint(:user)

Group Migration

defmodule XXXX.Repo.Migrations.CreateGroups do
  use Ecto.Migration

  def change do
    create table(:groups) do
      add :name, :string, null: false
      add :description, :text
      add :user_id, references(:users, on_delete: :delete_all), null: false




alias XXXX.Repo
alias XXXX.People.Group
alias XXXX.Accounts.User

user_example_1 = %User{} |> User.changeset(%{
  name: "Example User",
  username: "exampleuser",
  phonenumber: "1 834-000-0000",
  email: "",
  password: "secret"
}) |> Repo.insert!

example_group1 = %Group{} |> Group.changeset(%{
  name: "Family",
  description: "Random text",
  user: user_example_1
}) |> Repo.insert!

Hello, I am not sure I understand you use case: with this schema each groups record is related to only one users record (groups record belong to a single user, via the user_id foreign key): so you can have many groups records related to the same user, but you can’t have multiple users in the same group. Then I’m not sure this is what you want to model with your schema. In my opinion it would make more sense to relate users and groups with a many-to-many relationship, so you can have a user belonging to multiple groups and a group containing many different users. If that is the case, you can find more details in the ecto docs, else I apologize for misunderstanding your question.

Thanks for the response. I actually do want one user to have many groups. The groups are specific to each user and will be generated by the user. I really want to build a seed file passing the user struct to the creation of the group.

I see, :sweat_smile: sorry for the misleading answer then!

It is usually not a good idea to expose id field in the changeset.

You can do it in different ways, but if You have already a user, You could use put_assoc.

Another way to do it is to use build_assoc.

1 Like

Thank you! Reading through put_assoc I learned a lot more about associations and changesets. I can see why exposing the id field to the changeset is a bad idea, I will stick with using put and build for now. I appreciate your help again and patience, I am learning a lot!