show_all_values=T doen'st work for getDescriptionStatsBy

[peterolejua]

Hello,

I have some factors with some categories with 0 counts. How can I make describeFactors to show all values including 0's

Thanks

[gforge]

Could you write up a quick test and so I can have a look at this?

[gforge]

Wrote up a test and it works just as expected, note that descibeFactors is unaffected by this. It is assumed that dropping an empty factor is so easy that it is not the job of the function to clean empty values.

test_that("Factors with 0 observations should be reported when show_all_values=TRUE, issue #61", {
  set.seed(1)
  n = 20
  d <- data.frame(single = sample(LETTERS[1], size = n, replace = TRUE),
                  two = sample(LETTERS[1:2], size = n, replace = TRUE),
                  three = sample(LETTERS[1:3], size = n, replace = TRUE),
                  by = sample(letters[1:2], size = n, replace = TRUE)) %>%
    mutate(across(everything(), function(x) factor(x, levels = c(unique(x), "no observation"))))

  out <- d %>%
    getDescriptionStatsBy(single,
                          two,
                          three,
                          by = by,
                          show_all_values = TRUE)
  expect_true(all(sapply(out, function(x) "no observation" %in% rownames(x))))

  out <- d %>%
    getDescriptionStatsBy(single,
                          two,
                          three,
                          by = by,
                          show_all_values = FALSE)
  expect_false("no observation" %in% rownames(out[[1]]))
  expect_true(all(sapply(out[2:3], function(x) "no observation" %in% rownames(x))))
})