works just fine. The corruption looks a bit like a randomly scattered newline characters - the line suddenly ends (in the middle of a field), then there's a part of the original line missing and the remainder is on the next line. (or something like that, I'm not sure what is the exact corruption).
The data do not contain newline characters - when exporting to JSON everything is peachy. Also there are no mixed Windows/Unix line ends (I'm on Linux).