AlenkaF commented on a change in pull request #10519:
URL: https://github.com/apache/arrow/pull/10519#discussion_r652435928
##########
File path: r/tests/testthat/test-dplyr-arrange.R
##########
@@ -139,18 +139,14 @@ test_that("arrange() on integer, double, and character
columns", {
collect(),
tbl
)
- expect_warning(
- expect_equal(
- tbl %>%
- Table$create() %>%
- arrange(abs(int), dbl) %>%
- collect(),
- tbl %>%
- arrange(abs(int), dbl) %>%
- collect()
- ),
- "not supported in Arrow",
- fixed = TRUE
+ expect_equal(
Review comment:
Looking into `dplyr-arrange.R` I would agree that one test needs to be
left as there is a functionality in `arrange()` to check for things not
supported by Arrow and this seems to be the only case testing it.
https://github.com/apache/arrow/blob/f8661e032902a963b0a6a46077d72e804d22560d/r/R/dplyr-arrange.R#L39-L46
And so it doesn't make sense to keep `abs()` as it is now supported and it
was intended to test for `"try-error"` by `arrow_eval()`.
https://github.com/apache/arrow/blob/f8661e032902a963b0a6a46077d72e804d22560d/r/R/dplyr-eval.R#L18-L48
We could change `abs()` to for example `ceiling()`. But if it will ever be
supported in Arrow this will have to be changed again? And I am not sure how to
find a bulletproof function.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
