Improve list_simplify() performance + errors #942
Conversation
|
Oooh, and name management from https://github.com/tidyverse/purrr/pull/894/files#r957388812. |
R/list-simplify.R
Outdated
| x <- vec_set_names(x, NULL) | ||
| out <- list_unchop(x, ptype = ptype) |
There was a problem hiding this comment.
Hmm, I think this is slightly wrong because if the unchop fails and strict = FALSE, then we will return x which might have had its names stripped by this
Test for this case is probably useful, something like "returns a named list when input is named and simplification fails"
There was a problem hiding this comment.
Oops yes. Let me refactor.
R/list-simplify.R
Outdated
| @@ -83,8 +101,12 @@ simplify_impl <- function(x, | |||
| ) | |||
| } else { | |||
| if (strict) { | |||
| bad <- detect_index(x, function(x) vec_size(x) != 1) | |||
There was a problem hiding this comment.
You could also do
which(list_sizes(x) != 1L)[[1L]]or store the result of list_sizes(x) == 1L as where_size_one from when you computed it above in the strict branch along with the list_check_all_vectors() check, and then reference it here like bad <- which(!where_size_one)[[1L]]
There was a problem hiding this comment.
I moved this error up into the earlier strict branch which I think makes it easier to follow.
Co-authored-by: Davis Vaughan <davis@rstudio.com>
Fixes #935