nealrichardson commented on a change in pull request #9743:
URL: https://github.com/apache/arrow/pull/9743#discussion_r598970871
##########
File path: r/R/feather.R
##########
@@ -154,7 +154,10 @@ read_feather <- function(file, col_select = NULL,
as_data_frame = TRUE, ...) {
vars_select(names(reader), !!col_select)
}
- out <- reader$Read(columns)
+ out <- tryCatch(
+ reader$Read(columns),
+ error = function(e) { read_compressed_error(e) }
Review comment:
I don't see a `read_compressed_error()` function in this PR, and since
the tests seem to have passed, that means that none of the regular CI jobs
exercises this code path. So in addition to supplying the missing code, you
should think about which of our nightly test jobs would exercise this
condition. Indeed, you might want to run that job on this PR now, before you
fix the missing code, just to demonstrate that it is failing now. That way when
you run it after fixing your code, we'll see that the issue has been fixed
because the test passes.
##########
File path: r/tests/testthat/test-feather.R
##########
@@ -196,3 +196,19 @@ test_that("Character vectors > 2GB can write to feather", {
})
unlink(feather_file)
+
+ft_file <- system.file("v0.7.1.feather", package = "arrow")
+
+test_that("Error messages are shown when the compression algorithm lz4
+ is not found", {
+ skip_on_cran()
Review comment:
Why are these tests `skip_on_cran()`?
--
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]
