data.table vs dplyr: can one do something well the other can't or does poorly? data.table vs dplyr: can one do something well the other can't or does poorly? r r