Open hadley opened 6 years ago
Would it make sense for dots_names()
and dots_n()
to have an expand
argument? If TRUE
, it would expand !!!
to give the right answer when tidy dots are enabled. Should probably default to FALSE
because this can have side effects or be costly.
Maybe we can also check for !!!
when expand
is FALSE
and either return NA
or throw an error.
Or we could choose to make them blind to tidy dots. I think dots_n()
and dots_name()
are mostly used for generating error messages. It's not clear that they need to support tidy dots (or we could at least wait to implement until we have a concrete need)
hi! i'm writing a package that aims to help malaria metaanalysis researchers by allowing them to build tables declaratively, e.g.
MasterMalariaTableSchema <- BaseSchema(
ID=character,
temperature=numeric,
parasitaemia=numeric,
recrudescent=logical)
I was really struggling with ...
, but googling around and finding this page led to learning about rlang::exprs
, which I think is precisely what I need. Thanks <3
It's sometimes useful to capture the names of dotted arguments without evaluating them for generating error messages.
Related to
dots_n()