avoid with=FALSE in doc example

[MichaelChirico]

No description provided.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (cb8aeff) 97.49% compared to head (1813f67) 97.46%.
Report is 82 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #5594      +/-   ##
==========================================
- Coverage   97.49%   97.46%   -0.03%     
==========================================
  Files          80       80              
  Lines       14810    14822      +12     
==========================================
+ Hits        14439    14447       +8     
- Misses        371      375       +4     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[jangorecki]

I am not sure about this change.
For package developers who imports data.table the with=FALSE interface is generally preferred as they don't have to deal with extra NOTEs during package check about global variables. For the same reason, in my packages, I use DT[..., by=c("a","b")], rather than DT[..., by=.(a, b)].

[TysonStanley]

Since it is just an example, can we include both then? I think it makes sense to show both with something like:

To reorder \code{data.table} columns, the idiomatic way is to use \code{setcolorder(x, neworder)}, instead of doing \code{x <- x[, ..neworder]} (or \code{x <- x[, neworder, with=FALSE]}). This is because the latter makes an entire copy of the \code{data.table}, which maybe unnecessary in most situations. \code{setcolorder} also allows column numbers instead of names for \code{neworder} argument, although we recommend using names as a good programming practice.