two years ago, using keypaths in performance-sensitive code was not a good idea. the linked issue is still open.
has anything changed since?
I can't speak for the GitHub issue, but the benchmarks seem to have improved since I last tested (with [this repository](GitHub - atfelix/keypaths-performance: Key paths are slow and tested on Max M1 Max)):
name time std iterations
--------------------------------------------------------------------------------
-key path sorted :: name 12759325.500 ns ± 2.57 % 108
+key path sorted :: name 128666.000 ns ± 1.11 % 10823
-key path sorted :: age 14797394.000 ns ± 3.27 % 93
+key path sorted :: age 134375.000 ns ± 1.38 % 10371
-key path expressions sorted :: name 13089546.000 ns ± 3.94 % 106
+key path expressions sorted :: name 239833.000 ns ± 1.21 % 5817
-key path expressions sorted :: age 14440162.000 ns ± 2.42 % 95
+key path expressions sorted :: age 178208.000 ns ± 1.10 % 7815
-functions sorted :: name 102494.000 ns ± 9.12 % 13123
+functions sorted :: name 49041.000 ns ± 1.94 % 28500
-functions path sorted :: age 69360.000 ns ± 6.88 % 19774
+functions sorted :: age 41917.000 ns ± 4.10 % 32499
4 Likes