performance issue in will_make_matrix()

[simonpcouch]

The internal will_make_matrix() function lapply()s on an input that could be atomic:

parsnip/R/convert_data.R

Line 329 in 2588181

cls <- unique(unlist(lapply(y, class)))

This approach is quite a bit slower than something like check_outcome():

parsnip/R/misc.R

Lines 344 to 345 in 2588181

	outcome_is_factor <- if (is.atomic(y)) {is.factor(y)} else {all(map_lgl(y, is.factor))}
	if (!outcome_is_factor) {

Thankfully, most of the time, will_make_matrix() returns early before making it to this condition.

[simonpcouch]

Same goes for this vapply():

parsnip/R/convert_data.R

Lines 332 to 334 in 2588181

	can_convert <-
	vapply(y, function(x)
	is.atomic(x) & !is.factor(x), logical(1))

Pass factor(rep(0, 1e5)) to trigger both.

[simonpcouch]

With 81ad6b6 included as part of #888, I'm happy with where this is at now!

[github-actions]

This issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex: https://reprex.tidyverse.org) and link to this issue.