Bug: join_type = "before" produces confusing results

[bundfussr]

What happened?

If join_type = "before" is specified in derive_vars_joined() and dataset and dataset_add don't have the same observations, the result can be confusing.

Session Information

No response

Reproducible Example

library(tibble)
adbds <- tibble(subj = "1", day = c(15, 17, 20), val = c(-1, 0, 1))
derive_vars_joined(
  adbds,
  dataset_add = adbds %>% filter(val >= 0),
  by_vars = exprs(subj),
  order = exprs(day),
  new_vars = exprs(lastposval = val),
  join_type = "before",
  mode = "last"
)

produces

# A tibble: 3 × 4
  subj    day   val lastposval
  <chr> <dbl> <dbl>      <dbl>
1 1        15    -1         NA
2 1        17     0          0
3 1        20     1          1

instead of

# A tibble: 3 × 4
  subj    day   val lastposval
  <chr> <dbl> <dbl>      <dbl>
1 1        15    -1         NA
2 1        17     0         NA
3 1        20     1          0