scalar_in_linter() for x %in% 1

[MichaelChirico]

Part of #884

[AshesITR]

In the presence of NAs, x %in% "a" is much better than !is.na(x) & x == "a" IMO...

[AshesITR]

We could lint x %in% NA though to suggest is.na(x).

[MichaelChirico]

In the presence of NAs, x %in% "a" is much better than !is.na(x) & x == "a" IMO...

I prefer the latter for making missing behavior explicit... if you prefer the former it's best not to use this linter I'd say. We can tweak the documentation to reflect that it may be up to taste, WDYT?

[MichaelChirico]

We could lint x %in% NA though to suggest is.na(x).

I'd make that a message customization as a follow-up

[AshesITR]

In the presence of NAs, x %in% "a" is much better than !is.na(x) & x == "a" IMO...

I prefer the latter for making missing behavior explicit... if you prefer the former it's best not to use this linter I'd say. We can tweak the documentation to reflect that it may be up to taste, WDYT?

Fine by me, though I might turn on the RHS = NA case specifically.

[AshesITR]

We could lint x %in% NA though to suggest is.na(x).

I'd make that a message customization as a follow-up

Keep in mind that will generate incorrect advice since the behavior of == differs from %in% there.
It would be nice to restrict this linter to only this case btw because x %in% NA vs. is.na(x) is more clearly a lint IMO.
The logic should be quite simple - extract the constant as an R value and test it using is.na() or just hard-code the possible NAs, maybe offering capabilities for extension to NAs from other packages such as lubridate::NA_Date_.

[codecov-commenter]

Codecov Report

Merging #2084 (c3f43be) into main (35f3344) will increase coverage by 0.00%.
The diff coverage is 100.00%.

❗ Current head c3f43be differs from pull request most recent head 170d217. Consider uploading reports for the commit 170d217 to get more accurate results

@@           Coverage Diff           @@
##             main    #2084   +/-   ##
=======================================
  Coverage   99.65%   99.65%           
=======================================
  Files         118      119    +1     
  Lines        5210     5227   +17     
=======================================
+ Hits         5192     5209   +17     
  Misses         18       18           

Files Changed	Coverage Δ
R/scalar_in_linter.R	100.00% <100.00%> (ø)