GH-43440: [R] Unable to filter a factor column with %in% (#43446)

### Rationale for this change

Fixes #43440

### What changes are included in this PR?

The binding for `%in%` sends the DictionaryType's `value_type` to
`cast_or_parse()`. It's possible that it would be better to handle this
in `cast_or_parse()`, but it is used in lots of places and I wasn't sure
that was correct everywhere. We could certainly find out, but that's a
bigger testing exercise than I wanted to take on this afternoon.

### Are these changes tested?

Yes.

### Are there any user-facing changes?

The bug is fixed.
* GitHub Issue: #43440

r/R/dplyr-funcs-conditional.R

@@ -21,9 +21,16 @@ register_bindings_conditional <- function() {
     value_set <- Array$create(table)
     # If possible, `table` should be the same type as `x`
     # Try downcasting here; otherwise Acero may upcast x to table's type
+    x_type <- x$type()
+    # GH-43440: `is_in` doesn't want a DictionaryType in the value_set,
+    # so we'll cast to its value_type
+    # TODO: should this be pushed into cast_or_parse? Is this a bigger issue?
+    if (inherits(x_type, "DictionaryType")) {
+      x_type <- x_type$value_type
+    }
     try(
-      value_set <- cast_or_parse(value_set, x$type()),
-      silent = TRUE
+      value_set <- cast_or_parse(value_set, x_type),
+      silent = !getOption("arrow.debug", FALSE)
     )
 
     expr <- Expression$create("is_in", x,

r/tests/testthat/test-dplyr-funcs-conditional.R

@@ -26,6 +26,16 @@ tbl <- example_data
 tbl$verses <- verses[[1]]
 tbl$another_chr <- tail(letters, 10)
 
+test_that("%in% handles dictionary type", {
+  df <- tibble::tibble(x = factor(c("a", "b", "c")))
+  compare_dplyr_binding(
+    .input %>%
+      filter(x %in% "a") %>%
+      collect(),
+    df
+  )
+})
+
 test_that("if_else and ifelse", {
   compare_dplyr_binding(
     .input %>%