Ok, done. A clean design has been implemented for handling empty cells. Basically, empty input cells are ignored, and therefore they do not participate in the aggregation. (However, for aggregation using two columns (=corr(p,q)= for example), if a pair of cells contains both an empty and a non-empty cell, then the empty one is replaced by zero.)
On output, empty cells are generated when the aggregation function does not have enough input. For instance, =mean= needs at least one value, otherwise a division by zero happens. Thanks to Michael Brand for his insight and suggestions. Source code and documentation here: https://github.com/tbanel/orgaggregate Melpa installation: (require 'package) (add-to-list 'package-archives '("melpa" . "http://melpa.milkbox.net/packages";) t) (package-initialize) M-x package-list-packages install orgtbl-aggregate Have fun Thierry
