Recommend !is.unsorted() instead of identical(x, sort(x)) in sort_linter() #2908
Description
Not a super common pattern:
But there is a huge opportunity for optimization here:
x <- runif(1e5)
microbenchmark::microbenchmark(
identical(x, sort(x)),
!is.unsorted(x),
check = "identical"
)
#> Unit: microseconds
#> expr min lq mean median uq max
#> identical(x, sort(x)) 4010.383 4313.883 4636.98955 4496.855 4761.682 7040.232
#> !is.unsorted(x) 40.104 41.166 46.29157 42.666 46.807 86.509
#> neval
#> 100
#> 100
x <- sort(x)
microbenchmark::microbenchmark(
identical(x, sort(x)),
!is.unsorted(x),
check = "identical"
)
#> Unit: nanoseconds
#> expr min lq mean median uq max neval
#> identical(x, sort(x)) 3612 3732.0 4094.85 3872.5 4033.0 22994 100
#> !is.unsorted(x) 662 722.5 803.60 773.0 800.5 3131 100Created on 2025-07-30 with reprex v2.1.1
Should probably not lint if na.last = TRUE or FALSE or decreasing = TRUE in sort().