On 16 March 2021 at 10:50, Martin Maechler wrote: | I vaguely remember (from Luke's docs/presentation on ALTREP) | that there are some "missing parts" here. | One of them the not-existing R level functionality, another may be | the C code below R's is.unsorted() ... maybe is.unsorted() | could get a new argument and or be re-written, moving the NA | handling also to C and have that happen *after* the C code looks | if it's an ALTREP object and if that "knows it's sorted".
Somewhat related: I recently tried to demonstrate that, say, a 1e2 and a 1e6 vector each with one added NA should have equal "O(1)" time under anyNA(), yet apparently they do not---so the other magic bit (about knowing NAness) may not propagate. (That was with R 4.0.4.) I may also have missed a crucial bit about enabling ALTREPness here, yet the "gist of it" always was (AFAIK) that it is supposedly seamless. Dirk -- https://dirk.eddelbuettel.com | @eddelbuettel | e...@debian.org ______________________________________________ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel