ivoysey
commented
7 years ago i have an idea where we can make sure that we never misplace columns -- if you read the header file in order and have each row's entry be defined as the output of a function of some amount of the context (path, etc.) that defines the row that has the same name as the header in the file, you can just eval the functions in order and make it impossible to misplace data.
right now this is a bunch of hacks, probably not efficient, and not scalable. this is what we needed to do to be able to make data when we needed it. that said, this tool is probably something we want to use long term, when we can write our own tests and run then independently. that means it's probably worth it to tidy up the code base.