[R-Forge #5760] GForce should work with .N

[arunsrinivasan]

Submitted by: Arun ; Assigned to: Arun ; R-Forge link

At the moment, if one does:

DT[, list(mean(x), mean(y)), by=z]

It'll be optimised to use GForce. However, doing

DT[, list(.N, mean(x), mean(y)), by=z]

will result in dogroups instead of GForce, which could be slower. This case can be optimised relatively easily. Will do.