[Bug]: problem with datasets that have different merging keys with parent

[averissimo]

What happened?

This surfaced with insightsengineering/teal.data#215 (initially reported by @vedhav from CI fail in teal)

I believe it hasn't been caugh before as datasets usually have the same merging key

Example below uses a dataset from a test in teal.

td <- teal.data::teal_data() |> 
  within(d1 <- data.frame(id = 1:5, pk = c(2, 3, 2, 1, 4), val = 1:5)) |> 
  within(d2 <- data.frame(id = 1:5, value = 1:5))

teal.data::datanames(td) <- c("d1", "d2")
teal.data::join_keys(td) <- teal.data::join_keys(teal.data::join_key("d1", "d2", c("pk" = "id")))
teal.data::parents(td) <- list(d2 = "d1")

app <- teal::init(
  data = td,
  modules = list(
    teal::example_module(label = "example teal module")
  )
) |> shiny::runApp()

sessionInfo()

No response

Relevant log output

No response

Code of Conduct

• I agree to follow this project's Code of Conduct.

Contribution Guidelines

• I agree to follow this project's Contribution Guidelines.

Security Policy

• I agree to follow this project's Security Policy.

[averissimo]

The retrieval of parent_keys and dataset_keys needs correction. Diff below partially solves the issue (new error pops)

diff --git a/R/FilteredDatasetDataframe.R b/R/FilteredDatasetDataframe.R
index 06ca7511..c956ecf3 100644
--- a/R/FilteredDatasetDataframe.R
+++ b/R/FilteredDatasetDataframe.R
@@ -122,8 +122,8 @@ DataframeFilteredDataset <- R6::R6Class( # nolint
 
       if (!identical(parent_dataname, character(0))) {
         join_keys <- private$join_keys
-        parent_keys <- names(join_keys)
-        dataset_keys <- unname(join_keys)
+        parent_keys <- unname(join_keys)
+        dataset_keys <- names(join_keys)
 
         y_arg <- if (length(parent_keys) == 0L) {
           parent_dataname