Skip to content

ARROW-17738: [R] dplyr::compute should convert from grouped arrow_dplyr_query to arrow Table #14160

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
paleolimbot merged 12 commits into from
Oct 7, 2022
Merged

ARROW-17738: [R] dplyr::compute should convert from grouped arrow_dplyr_query to arrow Table #14160

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
12 commits
Select commit Hold shift + click to select a range
661cdc9
fix to return a Table
eitsupi Oct 1, 2022
4481819
collapse should return arrow dplyr query
eitsupi Oct 1, 2022
4b5f0c1
add tests for compute
eitsupi Oct 1, 2022
93f2f31
rename file to match name of file to be tested
eitsupi Oct 1, 2022
dc0ef4e
fix tests
eitsupi Oct 1, 2022
b2fbb15
remove unused line
eitsupi Oct 1, 2022
6231c17
remove unused line
eitsupi Oct 1, 2022
784f483
Revert "rename file to match name of file to be tested"
eitsupi Oct 1, 2022
42c1498
move test to the other file and rename the test case
eitsupi Oct 1, 2022
0aaee6f
Merge branch 'master' into r-fix-compute
eitsupi Oct 6, 2022
3ae0ca0
add tests for compute and collect
eitsupi Oct 6, 2022
47ec68d
more test
eitsupi Oct 6, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 5 additions & 4 deletions r/R/dplyr-collect.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -68,9 +68,7 @@ restore_dplyr_features <- function(df, query) {
)
} else {
# This is a Table, via compute() or collect(as_data_frame = FALSE)
df <- as_adq(df)
df$group_by_vars <- query$group_by_vars
df$drop_empty_groups <- query$drop_empty_groups
df$metadata$r$attributes$.group_vars <- query$group_by_vars
}
}
df
Expand All @@ -80,7 +78,10 @@ collapse.arrow_dplyr_query <- function(x, ...) {
# Figure out what schema will result from the query
x$schema <- implicit_schema(x)
# Nest inside a new arrow_dplyr_query (and keep groups)
restore_dplyr_features(arrow_dplyr_query(x), x)
out <- arrow_dplyr_query(x)
out$group_by_vars <- x$group_by_vars
out$drop_empty_groups <- x$drop_empty_groups
out
}
collapse.Dataset <- collapse.ArrowTabular <- collapse.RecordBatchReader <- function(x, ...) {
arrow_dplyr_query(x)
Expand Down
10 changes: 3 additions & 7 deletions r/tests/testthat/test-dataset-dplyr.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -284,13 +284,9 @@ test_that("compute()/collect(as_data_frame=FALSE)", {
group_by(fct) %>%
compute()

# the group_by() prevents compute() from returning a Table...
expect_s3_class(tab5, "arrow_dplyr_query")

# ... but $.data is a Table...
expect_r6_class(tab5$.data, "Table")
# ... and the mutate() was evaluated
expect_true("negint" %in% names(tab5$.data))
expect_r6_class(tab5, "Table")
# mutate() was evaluated
expect_true("negint" %in% names(tab5))
})

test_that("head/tail on query on dataset", {
Expand Down
44 changes: 42 additions & 2 deletions r/tests/testthat/test-dplyr-query.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -119,7 +119,7 @@ test_that("collect(as_data_frame=FALSE)", {
filter(int > 5) %>%
group_by(int) %>%
collect(as_data_frame = FALSE)
expect_s3_class(b4, "arrow_dplyr_query")
Copy link
Member

@nealrichardson nealrichardson Sep 20, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think we need a test (and matching code change) that a Table with groups in the $metadata prints that it is grouped. Currently, the print method does not reveal that you have groupings, even though they get restored when you as.data.frame():

> tab <- arrow_table(mtcars)
> tab$metadata$r$attributes$.group_vars <- "cyl"
> tab
Table
32 rows x 11 columns
$mpg <double>
$cyl <double>
$disp <double>
$hp <double>
$drat <double>
$wt <double>
$qsec <double>
$vs <double>
$am <double>
$gear <double>
$carb <double>

See $metadata for additional Schema metadata

I wonder what else is lost or concealed by this change. This feature of reading groups from $metadata$r$attributes$.group_vars was a bit of a hack and not something we're generally relying on, so I don't think it's a simple change. Or at least, it's not something we can just rely that the test suite already covers everything we care about. (Search for ".group_vars" in the code, it's not touched in many places, and only in one test.)

Not a reason not to do it, just need to be thorough.

Copy link
Contributor Author

@eitsupi eitsupi Sep 20, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for your review.

I think we need a test (and matching code change) that a Table with groups in the $metadata prints that it is grouped. Currently, the print method does not reveal that you have groupings, even though they get restored when you as.data.frame():

I agree that the print results need improvement.
But would it be better to create a separate ticket and do it there?
This occurs even in arrow 9.0.0, regardless of the behavior with compute.

mtcars |> dplyr::group_by(cyl) |> arrow::arrow_table()
#> Table
#> 32 rows x 11 columns
#> $mpg <double>
#> $cyl <double>
#> $disp <double>
#> $hp <double>
#> $drat <double>
#> $wt <double>
#> $qsec <double>
#> $vs <double>
#> $am <double>
#> $gear <double>
#> $carb <double>
#>
#> See $metadata for additional Schema metadata

Created on 2022-09-20 with reprex v2.0.2

This feature of reading groups from $metadata$r$attributes$.group_vars was a bit of a hack and not something we're generally relying on, so I don't think it's a simple change. Or at least, it's not something we can just rely that the test suite already covers everything we care about.

Since dplyr::group_vars reads this field, isn't this function generally used?

expect_r6_class(b4, "Table")
expect_equal(
as.data.frame(b4),
expected %>%
Expand Down Expand Up @@ -156,7 +156,7 @@ test_that("compute()", {
filter(int > 5) %>%
group_by(int) %>%
compute()
expect_s3_class(b4, "arrow_dplyr_query")
expect_r6_class(b4, "Table")
expect_equal(
as.data.frame(b4),
expected %>%
Expand Down Expand Up @@ -579,3 +579,43 @@ test_that("needs_projection unit tests", {
tab %>% relocate(lgl)
))
})

test_that("compute() on a grouped query returns a Table with groups in metadata", {
tab1 <- tbl %>%
arrow_table() %>%
group_by(int) %>%
compute()
expect_r6_class(tab1, "Table")
expect_equal(
as.data.frame(tab1),
tbl %>%
group_by(int)
)
expect_equal(
collect(tab1),
tbl %>%
group_by(int)
)
})

test_that("collect() is identical to compute() %>% collect()", {
tab1 <- tbl %>%
arrow_table()
adq1 <- tab1 %>%
group_by(int)

expect_equal(
tab1 %>%
compute() %>%
collect(),
tab1 %>%
collect()
)
expect_equal(
adq1 %>%
compute() %>%
collect(),
adq1 %>%
collect()
)
})