Just thought I’d mention that I recently added functionality for dealing with missing values to the data-frames library. It’s modelled mostly after R, so should be familiar. I like how the new data frame access methods are shaping up. Very lisp-like in the large. The docs are here:
in the end I went with a keyword sentinel value, :na, because using nil precluded having data-frames of Boolean values.
Comments and feedback always welcome.