Multiple Ecto transactions give timeout

Questions & Help Questions
, , ,
9

@dimitarvp it was inserting one by one. I changed it with the help of a friend and it worked. insert_all does the job as @axelson was saying. Now it looks like this:

def bulk_create(entity, items) do
    Ecto.Multi.new()
    |> Ecto.Multi.insert_all(:insert_all, entity, items, on_conflict: :nothing)
    |> Repo.transaction()
    |> case do
      {:ok, _} ->
        %{rows_affected: length(items), errors: []}

      {
        :error,
        _,
        %Ecto.Changeset{
          changes: changes,
          errors: errors
        },
        _
      } ->
        %{
          rows_affected: 0,
          errors: "Changes #{changes} couldn't be created because #{inspect(errors)}"
        }
    end
  end

But the problem is that I had to take care of timestamps because insert_all doesn’t insert them. So a migration to set a default value of now() solves the issue.

So, issue solved but I still won’t know why it was closing the connection even when I set the timeout to :infinity with the previous function. BTW, now with the insert_all solution I can stick to the default timeout of 15_000ms