Typo in vignette on reshaping

[unseldth]

I am currently working through the vignette on efficient reshaping and I think I came across one point, that is a bit unclear to me (or maybe wrong):
In section 3 (Enhanced (new) functionality), you say “We pass a list of columns to measure.vars”, but then your code passes a list to “measure” as “measure = list(colA, colB)”. To me, there is a difference between the measure() function, and the measure.vars argument of the melt function. I think, the code still works and the result is identical (probably, because the argument “measure” is automatically interpreted as “measure.vars”), but for clarity, I would change the melt function call to have the arguement “measure.vars = list(colA, colB)”, instead of “melt”, or am I misinterpreting something?

[ben-schwen]

The code works because of partial argument matching. This happens multiple times in the vignette and also at places before e.g.

dcast(DT.m1, family_id ~ ., fun.agg = function(x) sum(!is.na(x)), value.var = "dob")

vs.

args(dcast)
#> function (data, formula, fun.aggregate = NULL, ..., margins = NULL,
#>     subset = NULL, fill = NULL, value.var = guess(data))
#> NULL

You can find all of these occurences, via explicitly turning it off or warning about it with options(warnPartialMatchArgs=TRUE)

But I agree that we should probably eliminate partial matches for clarity in vignettes and examples.

[Mukulyadav2004]

Hi @ben-schwen
I'd like to work on this documentation issue to eliminate partial argument matching in vignettes and examples, as suggested. To identify occurrences, I plan to use options(warnPartialMatchArgs = TRUE) and systematically review the vignettes and test scripts for cases where partial matching happens (e.g., fun.agg instead of fun.aggregate).
Once identified, I'll update these instances with full argument names to improve clarity. Please let me know if I may start working on this.
Thanks!

[MichaelChirico]

@Mukulyadav2004 that sounds great, thanks.