Can't we drastically reduce the amount of purrr variants, AND expand possibilities, by using adverbs ?

[moodymudskipper]

We would be less efficient but can't we create any purrr::map* function from lapply by using adverbs/tags/adverb factories for :

• use of formula lambdas : map etc
• type stability : map_dbl, map_chr , modify etc
• using index or name : imap etc
• l* variants : lmap etc
• *_at / *_if variants : map_if, map_at etc
• _depth variants : map_depth, modify_depth
• special variants : map_if_else (upcoming), map_case_when (requested)
• map2 variants : map2; map2_chr etc
• pmap variants : pmap, pmap_chr etc

The way purrr seems to work now is that variants are created for each of those, AND if a combination is deemed common enough it gets its own verb. Inevitably we sometimes want a combination that doesn't exist, but they can't all exist.

The goal is not to replace the existing variants, but by extracting these features into independant adverbs (tags, adverb factories...) we can :

• fill the gaps in purrr
• Use these features on functions from other packages

Type stability adverbs are really simple to implement as it's just post processing.

We already implemented the lambda formula adverb as a simple preprocessing on function arguments using heuristics (but using adverb factories we could be more flexible and name the argument to preprocess).

Adverbs for other variants might be more complex and less easy to generalize to a wide range of input functions but it's worth taking a shot.

This family of adverbs might be a nice use case for dialects , we could throw in an adverb for progress bars, and quasi quotation too.

[moodymudskipper]

to the fridge for now, the most useful variants would be to build the functionalities of the i prefix and _at _if suffixes