I've loved the idea of Odo since I first saw it, but I've always been wary: the devil is in the details. I'm curious when and how information gets lost during e

I've loved the idea of Odo since I first saw it, but I've always been wary: the devil is in the details. I'm curious when and how information gets lost during each transfer because of the peculiarities of each format, or how those decisions are made and exposed. Min/max limits, truncation nulls, floating point precision, encodings, picking CHAR vs VARCHAR or string vs categorical, metadata like indices, etc are some of the hard problems behind bulk loading.