Some tests failing with the upcoming release of rnaturalearthdata

[PMassicotte]

Hi.

We are planing to release rnaturlearthdata v1.0.0 which ends support to sp in favour of sf.

It appears that these two tests are failing:

BeeBDC/tests/testthat/test-countryOutlieRs.R

Lines 32 to 40 in 2014ec3

	testthat::test_that("countryOutlieRs results TRUE/passed", {
	testthat::expect_equal(sum(testOut$.countryOutlier == TRUE, na.rm = TRUE), 74)
	})
	testthat::test_that("countryOutlieRs results FALSE/failed", {
	testthat::expect_equal(sum(testOut$.countryOutlier == FALSE, na.rm = TRUE), 3)
	})
	testthat::test_that("countryOutlieRs results NA/could not assess", {
	testthat::expect_equal(sum(is.na(testOut$.countryOutlier)), 23)
	})

Can you confirm on your side?

Regards,
Phil

[jbdorey]

Hi Phil, Thanks for letting us know! I’ve had a little look and I think that the problem on the BeeBDC side is that the “iso_a3” column now returns -99 for (at least for this test) Norway and France. As do a few other columns. So, my question would be — should I be looking to use another column, or is this a data issue on your end? I know that country names and codes are a real pain, so hopefully this isn’t too difficult :) Warm regards, James 

[jbdorey]

And sorry, that is from using this code:countryMap <- rnaturalearth::ne_countries(returnclass = "sf", country = NULL,                                         type = "countries", scale = 50)  %>%                                             # buffer by zero and make geometry valid to avoid potential issues with polygon intersection                                           sf::st_make_valid() %>%                                         # Select only a subset of the naturalearthdata columns to extract                                         dplyr::select(iso_a2, iso_a3, name, name_long, continent, geometry) %>%                                         # Replace some country codes to match the checklist                                         dplyr::mutate(iso_a3 = iso_a3 %>%                                        stringr::str_replace_all(    c("ALA" = "FIN", # Aland Islands == Finland    "ASM" = "WSM", # American Samoa == Samoa    "HKG" = "CHN", # Hong Kong == China    "MAF" = "SXM"))) # Saint-Martin == Sint Maarten

[PMassicotte]

Hi @jbdorey !

Do I understand that

countryMap <- rnaturalearth::ne_countries(
  returnclass = "sf",
  country = NULL,
  type = "countries",
  scale = 50
) %>%
  sf::st_make_valid() %>%
  dplyr::select(iso_a2, iso_a3, name, name_long, continent, geometry) %>%
  dplyr::mutate(iso_a3 = iso_a3 %>% stringr::str_replace_all(
    c(
      "ALA" = "FIN",
      "ASM" = "WSM",
      "HKG" = "CHN",
      "MAF" = "SXM"
    )
  ))

Would fix the issue?

And congratulation for your new job!

[PMassicotte]

Looks like that in the new version of the data, iso_a3_eh contains the info:

library(rnaturalearthdata)
countries110[countries110$iso_a3_eh == "FRA"]
countries110[countries110$iso_a3_eh == "NOR"]

r$> setdiff(countries110$iso_a3_eh, countries110$iso_a3)
[1] "NOR" "FRA"

[PMassicotte]

Might have some info here:

• Norway ISO_A3_EH Code is -99 nvkelso/natural-earth-vector#252
• France showing as -99 for ISO_A2 and ISO_A3 nvkelso/natural-earth-vector#695

[jbdorey]

Hey Phil, Thanks for that detective work, indeed that has fixed up a few of these new -99 cells! Many thanks for your detective work! When are you planning your release? I will be starting work next week and high on my to-do-list is to run BeeBDC on an ARM mac and potentially a windows emulator using   work computers (although I think it still pass all tests without). Then I will try to push 1.0.4 as close to your release as I can!  Many thanks, James 

[PMassicotte]

Thank you very much!

I will submit later today :)