[jira] [Updated] (ARROW-14908) [R] join followed by a summarize and collect crashes R

["Fabio Machado (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/ARROW-14908?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Fabio Machado updated ARROW-14908:
----------------------------------
    Description: 
{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}

  was:
{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 ------------------------------------------------------------------
#crashes RStudio
other_mtcars_data %>% 
  left_join(cars_arrow) %>% 
  count(main_color) %>% 
  collect()
#other variation also crash RStudio
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}


> [R] join followed by a summarize and collect crashes R
> ------------------------------------------------------
>
>                 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
>            Priority: Critical
>
> {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)