If you want to check that a changeset is valid in a function head you normally do it something like:

def some_func(%Ecto.Changeset{valid?: true} = changeset) do
  ...
end

as this allows you to bind a name to the changeset and also perform the validation check in a concise way. You can also bind nested vars this way. e.g.

def some_func(%Ecto.Changeset{valid?: true, changes: %{password: pass}}) do
  ..
end

If you just want to check if a changeset is valid and you are not pattern matching, then yes, you should just check changeset.valid?.

You should be able to access the properties with a dot notation. For example look here:

http://blog.danielberkompas.com/elixir/2015/05/08/testing-ecto-validations.html

def create(conn, params) do
  user = MyApp.User.changeset(%MyApp.User{}, params)

  if user.valid?
    MyApp.Repo.insert(user)
    # return success
  else
    # render error
  end
end

Do you do something similar?

cool man!

no, in my case on the phoenix book i bought, it does something like %User{“is_valid” => return_value} = changeset, something like that, and grab that value from the struct.

The point is, i understand that is maybe the best thing to do pattern matching, but it seem like i’m forced to use it instead of the dot syntax.

So if there is no particular reason, is ok if i use the dot syntax?

thanks.