[
https://issues.apache.org/jira/browse/ARROW-14908?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17489860#comment-17489860
]
Will Jones commented on ARROW-14908:
------------------------------------
I've made some progress, but not yet found a fix. This issue isn't strictly
linked to the Windows platform; instead it has do to with the setting
{{options(arrow.use_threads = FALSE)}}, which it doesn't seem the join
implementation anticipated. That setting is FALSE on Windows by default, but
TRUE on other platforms. I can reproduce this issue on Mac OS by setting it to
FALSE.
Note that I can get the example to work on Windows by setting
{{options(arrow.use_threads = TRUE)}}, so that may be a viable workaround in
some cases. However, other issues have been reported on Windows with the
use_threads option, which is why it has been turned off for now on Windows.
> [R] join on dataset crashes on Windows
> --------------------------------------
>
> Key: ARROW-14908
> URL: https://issues.apache.org/jira/browse/ARROW-14908
> Project: Apache Arrow
> Issue Type: Bug
> Components: R
> Affects Versions: 6.0.0
> Environment: R version 4.0.4
> Reporter: Fabio Machado
> Assignee: Will Jones
> Priority: Critical
> Labels: pull-request-available
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> {code:java}
> library(tidyverse)
> library(arrow)
> car_info <- rownames_to_column(mtcars, "car_info")
> cars_arrow_table <- arrow_table(car_info)
> other_mtcars_data <- select(car_info, 1) %>%
> mutate(main_color = sample( c("red", "blue", "white", "black"), size = n(),
> replace = TRUE)) %>%
> arrow::arrow_table()
> temp <- tempdir()
> par_temp <- paste0(temp, "\\parquet")
> car_info %>% arrow::write_dataset(par_temp)
> cars_arrow <- arrow::open_dataset(par_temp)
> # using arrow tables works
> ------------------------------------------------------
> cars_arrow_table %>% left_join(other_mtcars_data) %>% count(main_color) %>%
> collect()
> # using open dataset crashes R
> ------------------------------------------------------------------
> other_mtcars_data %>%
> left_join(cars_arrow) %>%
> count(main_color) %>%
> collect()
> #other variation also crash
> cars_arrow %>%
> left_join(other_mtcars_data) %>%
> count(main_color) %>%
> collect()
> cars_arrow %>%
> left_join(other_mtcars_data) %>%
> group_by(main_color) %>%
> summarise(n = n()) %>%
> collect()
> #compute also crashes
> cars_arrow %>%
> left_join(other_mtcars_data) %>%
> count(main_color) %>%
> compute()
> # workaround with duckdb
> ------------------------------------------------------
> ##this works
> cars_duck <- to_duckdb(cars_arrow, auto_disconnect = TRUE)
> other_cars_duck <- to_duckdb(other_mtcars_data, auto_disconnect = TRUE)
>
> cars_duck %>%
> left_join(other_cars_duck) %>%
> count(main_color) %>%
> collect()
> ##this doesn't (don't know if expected to work actually)
> cars_arrow %>%
> left_join(other_mtcars_data) %>%
> to_duckdb() {code}
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
