I am using a CSV dataset in SQLite to seed. It seems that I can insert data but I can’t read it ! Something obvious must be missing but I just don’t see.

The table looks like this:

def create_table(conn, table) do
    Exqlite.Sqlite3.execute(
      conn,
      "CREATE TABLE IF NOT EXISTS #{table} (
        id integer primary key,
        ident string,
        elevation_ft string,
        type string,
        name string,...

I also have a schema and a Repo.

defmodule Airport do
  use Ecto.Schema
  import Ecto.Changeset
  @headers headers
  @columns Enum.map(@headers, &String.to_atom/1)

  schema "csv" do
    Enum.each(@columns, &field(&1, :string))
    #timestamps()
  end

  def changeset(attrs \\ %{}) do
    %Airport{}
    |> cast(attrs, @columns)
    |> validate_required(@columns -- [:iata_code])
  end
end

I check that the fields are :ok

Airport.__schema__(:fields)
[:id, :ident, :type, :name, :elevation_ft...]

Now it seems that I can insert (?) a row:

l = %{
      coordinates: "-74.93360137939453, 40.07080078125",
      elevation_ft: "11",
      gps_code: "00A",
      ...
}
|> Airport.changeset()
|> Repo.insert()

which returns:

[debug] QUERY OK db=0.9ms idle=1248.3ms
INSERT INTO "csv" (coordinates","elevation_ft","gps_code....

{:ok,
 %Airport{
   __meta__: #Ecto.Schema.Metadata<:loaded, "csv">,
   id: 57424,
   ident: "00A",....

but I can’t read it !??

Repo.get(Airport, 1)
** (ArgumentError) cannot load `11` as type :string for field :elevation_ft in %Airport{__meta__:

It seems that I can insert my whole file, but I can’t read it either:

Repo.aggregate(Airport, :count)
57342

Seems good, but:

Repo.all(Airport)
** (ArgumentError) cannot load `11` as type :string for field :elevation_ft in 
%Airport{__meta__: #Ecto.Schema.Metadata<:loaded, "csv">, id: nil, ident: nil, type: nil, name: nil, elevation_ft: nil,
...