[bug] joins with incorrect names should error out, not OUTER JOIN.

[daikonradish]

Pretty sure joins are buggy.

1. Load up Titanic dataset, call itdf
2. Try the following: df |> rightJoin ["Cats"] df
3. Expected: Column not found "Cats". But instead, got a successful join with 793881 rows. (This is equal to 891 * 891, with 891 being the number of rows in Titanic)
4. This leads me to believe that the columns are being full outer joined if the column doesn't exist.

PS - I think that this should be a beginner-friendly task so please farm it out!