[R-Forge #1974] Update wiki for j=lapply(.SD,...) speedup in v1.8.1. Allow .SDcols to accept unquoted names.

[arunsrinivasan]

Submitted by: Yike Lu; Assigned to: Nobody; R-Forge link

See Dos/Don'ts number 1:

http://rwiki.sciviews.org/doku.php?id=packages:cran:data.table

Essentially, it would be nice for ans2 to be the same as ans1 without colnames(ans2)=colnames(ans1)

[jaapwalhout]

I think it would be consistent behavior if .SDcols also accepts lists as follows:

DT[, lapply(.SD, function), by = .(id, group), .SDcols = .(a, b, c)]

Related to #1257

[DavidArenburg]

I agree with @jaapwalhout, never noticed this inconsistency before.

[franknarf1]

Re .SDcols= accepting unquoted names, if the extension is general enough, maybe it could also be used in other places where cols= appear, like na.omit(x, cols = DC3:DC9) from SO: https://stackoverflow.com/q/51385594/

[MichaelChirico]

I think this is out-of-scope, maintainers don't want to add more NSE to .SDcols, though it could support more other features.

As to the "do's/dont's" in the OP, that link is dead, so context is lost. A new issue should help modernize the request if still relevant.