Jan Gorecki

one more approach not listed here is to pass `substitute()` to eval-call ```r my_sample

And what about C?

you might find following discussion related: https://github.com/Rdatatable/data.table/issues/2422

I would rather prefer to change `cbind.data.frame` to match `rbind.data.frame` behavior.

Yes the installation steps are broken, the easiest way to reproduce is just start clean Ubuntu 16.04 and go step by step with HOWTO doc. I was able to fix...

Benchmarking SAS might not be feasible, license of most of closed source solutions prohibits publishing any benchmarks.

@simison separate tab for message that you are not interested to browse through anymore. Same as on CS.

Yes

I cannot remove the new label unfortunatelly, can we re-open this?

The R data.table implementation (as of now rollmean) was designed to be R agnostic, so we should be able to reduce maintanence effort by reusing that code.