Try/except block in outer joins isn't atomic about csvkit HOT 1 CLOSED

You're right, I meant to fix this but couldn't quite come up another formulation I liked. That for...in should be moved outside the try block. It doesn't blow up in any tests because the logic in that loop is fairly trivial, but that's not really an excuse for potentially silencing an unexpected error.

Regarding abstracting an _outer_join construct: yes, absolutely. I was tempted to do it with metaclasses last night, then decided that was probably a pretty good indication I should go to bed.

Re: one-to-many's, unless your sold on the logic I'd like to talk this one through. I was thinking about it tonight and realized that I'm not actually sure I even know what the canonical behavior is in cases like that! If it is what I think it is, then its going to be tricky to implement. Either way, it would be good to clarify.

from csvkit.