Hey... Just wondering what core.matrix
expects as column names these days? Should they be strings or keywords? Is one enforced, and are they normalised at query time like incanter does?
One problem I have with incanter is that it allows both strings and keywords as column names - and then it cananonicalises on its API operations - however this causes lots of problems as two datasets which a user expects to be =
might be not=
etc... wondering whether core.matrix does anything in this regard?