This is an automated email from the ASF dual-hosted git repository.
jonkeane pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/arrow.git
The following commit(s) were added to refs/heads/main by this push:
new 1e8645e803 GH-48660: [R] Add tests for filter() and arrange() with
aggregation expressions (#48661)
1e8645e803 is described below
commit 1e8645e803f4861dfa869725c9b3426f1afbefca
Author: Hyukjin Kwon <[email protected]>
AuthorDate: Sun Jan 11 12:19:15 2026 +0900
GH-48660: [R] Add tests for filter() and arrange() with aggregation
expressions (#48661)
### Rationale for this change
03f8ae754e added a couple of TODOs for testing aggregation expressions
https://github.com/apache/arrow/blob/03f8ae754ede16f118ccdba0abb593b1461024aa/r/R/dplyr-filter.R#L49
https://github.com/apache/arrow/blob/03f8ae754ede16f118ccdba0abb593b1461024aa/r/R/dplyr-arrange.R#L54
Should test negative cases too as mentioned in the TODOs.
### What changes are included in this PR?
This PP adds a couple of tests for filter() and arrange() with
aggregation expressions
### Are these changes tested?
Unittests added.
### Are there any user-facing changes?
No, test-only.
* GitHub Issue: #48660
---
r/R/dplyr-arrange.R | 3 +--
r/R/dplyr-filter.R | 1 -
r/tests/testthat/test-dplyr-arrange.R | 8 ++++++++
r/tests/testthat/test-dplyr-filter.R | 8 ++++++++
4 files changed, 17 insertions(+), 3 deletions(-)
diff --git a/r/R/dplyr-arrange.R b/r/R/dplyr-arrange.R
index 977ccf9860..5394eeb539 100644
--- a/r/R/dplyr-arrange.R
+++ b/r/R/dplyr-arrange.R
@@ -46,10 +46,9 @@ arrange.arrow_dplyr_query <- function(.data, ..., .by_group
= FALSE) {
# dplyr lets you arrange on e.g. x < mean(x), but we haven't
implemented it.
# But we could, the same way it works in mutate() via join, if someone
asks.
# Until then, just error.
- # TODO: add a test for this
arrow_not_supported(
.actual_msg = "Expression not supported in arrange() in Arrow",
- call = expr
+ call = exprs[[i]]
)
}
descs[i] <- x[["desc"]]
diff --git a/r/R/dplyr-filter.R b/r/R/dplyr-filter.R
index 6f64749bbc..18f5c929af 100644
--- a/r/R/dplyr-filter.R
+++ b/r/R/dplyr-filter.R
@@ -42,7 +42,6 @@ filter.arrow_dplyr_query <- function(.data, ..., .by = NULL,
.preserve = FALSE)
# dplyr lets you filter on e.g. x < mean(x), but we haven't
implemented it.
# But we could, the same way it works in mutate() via join, if someone
asks.
# Until then, just error.
- # TODO: add a test for this
arrow_not_supported(
.actual_msg = "Expression not supported in filter() in Arrow",
call = expr
diff --git a/r/tests/testthat/test-dplyr-arrange.R
b/r/tests/testthat/test-dplyr-arrange.R
index c0e2b859c1..d502a7f040 100644
--- a/r/tests/testthat/test-dplyr-arrange.R
+++ b/r/tests/testthat/test-dplyr-arrange.R
@@ -245,3 +245,11 @@ test_that("Can use across() within arrange()", {
example_data
)
})
+
+test_that("arrange() with aggregation expressions errors", {
+ tab <- arrow_table(tbl)
+ expect_warning(
+ tab |> arrange(int < mean(int)),
+ "not supported in arrange"
+ )
+})
diff --git a/r/tests/testthat/test-dplyr-filter.R
b/r/tests/testthat/test-dplyr-filter.R
index 23d49feb3b..d56e25fca3 100644
--- a/r/tests/testthat/test-dplyr-filter.R
+++ b/r/tests/testthat/test-dplyr-filter.R
@@ -490,3 +490,11 @@ test_that(".by argument", {
"Can't supply `\\.by` when `\\.data` is grouped data"
)
})
+
+test_that("filter() with aggregation expressions errors", {
+ tab <- arrow_table(tbl)
+ expect_warning(
+ tab |> filter(int < mean(int)),
+ "not supported in filter"
+ )
+})
