test_that("error handling works", {
  chr <- letters[1:10]
  lgc <- rep(c(TRUE, FALSE), 5)
  nmr <- 1:10
  nmr2 <- 1

  expect_error(
    percent_change(chr, nmr),
    "`baseline` must be numeric; not character.",
    class = "error_argument_type"
  )
  expect_error(
    percent_change(lgc, nmr),
    "`baseline` must be numeric; not logical.",
    class = "error_argument_type"
  )
  expect_error(
    percent_change(nmr, chr),
    "`followup` must be numeric; not character.",
    class = "error_argument_type"
  )
  expect_error(
    percent_change(nmr, lgc),
    "`followup` must be numeric; not logical.",
    class = "error_argument_type"
  )
  expect_error(
    percent_change(nmr, nmr2),
    "`baseline` and `followup` must have the same length.",
    class = "error_argument_diff_length"
  )
})

test_that("percent_change() works", {
  set.seed(20200525)

  change <- c(
    1, 1.005, 1.05, 1.5, 2,
    0.995, 0.95, 0.5, 0, 1
  )
  baseline <- sample(1:10)
  baseline[10] <- NA
  followup <- baseline * change

  out <- percent_change(baseline, followup)

  expect_s3_class(out, c("lvmisc_percent", "vctrs_vctr"), exact = TRUE)

  expected <- vec_c(0, 0.005, 0.05, 0.5, 1, - 0.005, - 0.05, - 0.5, - 1, NA)
  expected <- as_percent(expected)

  expect_equal(out, expected)
})