IDE/ R session crashed with read_delim/mutate

[rg4555]

Hello,

Quite a few months ago I wrote this code for importing a csv file that worked fine at the time :

library(tidyverse)
###################################################################################### p 1000 t1 #################
# read each row of text individually so we can parse out the information manually
election0 <- 
  read_delim(
    "~/MASTER_1/stage/travail/R/donnees/2014/2014_t1+1000.txt", 
    "\n",
    col_names = FALSE,locale = locale(encoding = "ISO-8859-1")) %>%
  setNames("line_text") %>%
  mutate(
    # split by delimiter
    split_text  = strsplit(line_text, ";"),
    # assume the first 17 elements are common
    split_df    = map(split_text, ~.[1:17]),
    # and everything past this is repeating 11
    split_names = map(split_text, ~.[-c(1:17)]),
    columns     = map_dbl(split_text, length),
    # the number of repeating 11 name data elements
    n_names     = (columns - 17)/11) 

But when I run this same code on the same data as before my IDE crash (I think) I get a dialog box saying
R session aborded, R encountered a fatal error. The session was terminated :

Here is my session info :

> sessionInfo()
R version 4.1.0 (2021-05-18)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 19043)

Matrix products: default

locale:
[1] LC_COLLATE=French_France.1252  LC_CTYPE=French_France.1252   
[3] LC_MONETARY=French_France.1252 LC_NUMERIC=C                  
[5] LC_TIME=French_France.1252    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] forcats_0.5.1   stringr_1.4.0   dplyr_1.0.7     purrr_0.3.4     readr_2.0.0    
[6] tidyr_1.1.3     tibble_3.1.3    ggplot2_3.3.5   tidyverse_1.3.1

loaded via a namespace (and not attached):
 [1] Rcpp_1.0.7       cellranger_1.1.0 pillar_1.6.2     compiler_4.1.0   dbplyr_2.1.1    
 [6] tools_4.1.0      lubridate_1.7.10 jsonlite_1.7.2   lifecycle_1.0.0  gtable_0.3.0    
[11] pkgconfig_2.0.3  rlang_0.4.11     reprex_2.0.1     cli_3.0.1        rstudioapi_0.13 
[16] DBI_1.1.1        haven_2.4.3      xml2_1.3.2       withr_2.4.2      httr_1.4.2      
[21] fs_1.5.0         generics_0.1.0   vctrs_0.3.8      hms_1.1.0        grid_4.1.0      
[26] tidyselect_1.1.1 glue_1.4.2       R6_2.5.0         fansi_0.5.0      readxl_1.3.1    
[31] tzdb_0.1.2       modelr_0.1.8     magrittr_2.0.1   backports_1.2.1  scales_1.1.1    
[36] ellipsis_0.3.2   rvest_1.0.1      assertthat_0.2.1 colorspace_2.0-2 utf8_1.2.2      
[41] stringi_1.7.3    munsell_0.5.0    broom_0.7.9      crayon_1.4.1    

And an extract of the logfile :

09 Aug 2021 07:03:02 [rsession-zugat] ERROR system error 10053 (Une connexion établie a été abandonnée par un logiciel de votre ordinateur hôte) [request-uri: /events/get_events]; OCCURRED AT void __cdecl rstudio::session::HttpConnectionImpl<class rstudio_boost::asio::ip::tcp>::sendResponse(const class rstudio::core::http::Response &) src/cpp/session/http/SessionWin32HttpConnectionListener.cpp:113; LOGGED FROM: void __cdecl rstudio::session::HttpConnectionImpl<class rstudio_boost::asio::ip::tcp>::sendResponse(const class rstudio::core::http::Response &) src/cpp/session/http/SessionWin32HttpConnectionListener.cpp:118
09 Aug 2021 07:18:18 [rsession-zugat] ERROR system error 5 (Accès refusé); OCCURRED AT auto __cdecl rstudio::core::system::ChildProcess::terminate::<lambda_b34d56978c1a268cda78ea8a24bc0d35>::operator ()(void) const src/cpp/core/system/Win32ChildProcess.cpp:287; LOGGED FROM: void __cdecl rstudio::core::system::ProcessSupervisor::terminateAll(void) src/cpp/core/system/Process.cpp:363

It's the first time I see the second error since I'm trying to execute the code so i don't think that one is causing trouble but I don't know.

Also with the help of rstudio community, it seems that it's the read_delim function that caused issue since we were able to import the file ussing read_lines.

Note : the data and code has no change since the time it was working fine.

plus a link to the thread : https://community.rstudio.com/t/ide-r-session-crashed-with-mutate/112330

[romainfrancois]

I'm not sure there is much we can do at the dplyr level. Can you try to un-pipe the example so that at least you know where it happens.

Did you try outside of the IDE ? Perhaps it is an IDE issue ?

Unfortunately, without a usable reprex, there isn't much more we can do to help, e.g. we don't have "~/MASTER_1/stage/travail/R/donnees/2014/2014_t1+1000.txt"

[rg4555]

So the first things I did was to un-pipe it and it worked fine until the mutate segment.
But I don't think that the mutate function is the cause of it since when i'm scouting the data with view() it crash the session as well. also it worked using read_lines() instead.

I tried with rgui but it crashed too, besides it says r session crashed and not the IDE. I think that if the IDE crashed it wouldn't prompt anything.

unfortunately i can't provide a usefull reprex since the r session crashes and reprex() isn't working with a code that run into a crash.

the data are available here on the french government site :
https://www.data.gouv.fr/fr/datasets/elections-municipales-2014-resultats-1er-tour/

[DavisVaughan]

This seems like a potential readr/vroom bug (maybe ALTREP related). What happens if you try readr::read_delim(lazy = FALSE) when reading the file?

[rg4555]

Still crashing...

[DavisVaughan]

I can reproduce this as a vroom bug, and opened an issue over there. It seems to be related to guess_max

[jimhester]

The bottom line is that '\n' is not a valid delimiter to use in read_delim(). For CSV and other delimited files newlines are always record delimiters, never field delimiters.

This worked by accident in older version of readr, but it was never the intended usage.

I have added a workaround to avoid this problem in vroom in the future.

However if your intent is to just read the lines into a vector I would suggest using readr::read_lines() rather than read_delim().