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

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

Fabio Machado created ARROW-14908:
-------------------------------------

             Summary: [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


{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}



--
This message was sent by Atlassian Jira
(v8.20.1#820001)