Confusing error message if type differs between units objects

[kendonB]

See below:

library(dplyr)
library(units)
#> udunits system database from C:/R/Library/3.5/units/share/udunits
if_else(TRUE, c(1) %>% set_units("m"), c(1L) %>% set_units("m"))
#> Error: `false` must be a `units` object, not a `units` object

^{Created on 2019-01-30 by the reprex package (v0.2.1.9000)}

[romainfrancois]

🤔 so this happens in check_type():

check_type <- function(x, template, header) {
  if (identical(typeof(x), typeof(template))) {
    return()
  }

  glubort(header, "must be {friendly_type_of(template)}, not {friendly_type_of(x)}")
}

where it checks the underlying type but then errors based on a higher level class. It might be something vctrs could help with, perhaps with implementations of vctrs::vec_ptype_abbr() and/or vctrs::vec_ptype_full() for units objects.

[hadley]

This will be fixed through a complete overhaul of ifelse

[kendonB]

This is also a problem with case_when, fyi.

[hadley]

Yeah, it's a problem with a bunch of functions. We know about it and we have a plan to fix it.

[hadley]

Duplicate of #2432

[lock]

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