“Computation failed in stat_count() when using loadRDS

[nick-youngblut]

I often save my ggplot objects with saveRDS() so that I can load & plot them latter in new sessions. This works quite well, except when I use geom_bar(), which creates the following warning:

Warning message:
“Computation failed in `stat_count()`:
Please use `data_frame()` or `new_data_frame()` instead of `data.frame()` for better performance. See the vignette "ggplot2 internal programming guidelines" for details.”

...and results in an empty plot.

Reprex:

Session 1:

library(dplyr)
library(ggplot2)
out_file = 'ggplot_obj.RDS'
mpg %>%
    ggplot(aes(class)) +
    geom_bar() %>%
    saveRDS(file=out_file)

Session 2:

library(dplyr)
library(ggplot2)
out_file = 'ggplot_obj.RDS'
p = loadRDS(out_file)
plot(p)

sessionInfo

R version 3.6.2 (2019-12-12)
Platform: x86_64-conda_cos6-linux-gnu (64-bit)
Running under: Ubuntu 18.04.3 LTS

Matrix products: default
BLAS/LAPACK: /ebio/abt3_projects/Georg_animal_feces/envs/tidyverse/lib/libopenblasp-r0.3.7.so

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
 [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

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

other attached packages:
[1] RColorBrewer_1.1-2 ggplot2_3.2.1      tidyr_1.0.0        dplyr_0.8.3       

loaded via a namespace (and not attached):
 [1] Rcpp_1.0.3       plyr_1.8.5       pillar_1.4.3     compiler_3.6.2  
 [5] base64enc_0.1-3  tools_3.6.2      zeallot_0.1.0    digest_0.6.23   
 [9] uuid_0.1-2       jsonlite_1.6     evaluate_0.14    tibble_2.1.3    
[13] lifecycle_0.1.0  gtable_0.3.0     pkgconfig_2.0.3  rlang_0.4.2     
[17] IRdisplay_0.7.0  IRkernel_1.1     stringr_1.4.0    repr_1.0.2      
[21] withr_2.1.2      vctrs_0.2.1      grid_3.6.2       tidyselect_0.2.5
[25] glue_1.3.1       R6_2.4.1         pbdZMQ_0.3-3     purrr_0.3.3     
[29] farver_2.0.2     reshape2_1.4.3   magrittr_1.5     backports_1.1.5 
[33] scales_1.1.0     htmltools_0.4.0  assertthat_0.2.1 colorspace_1.4-1
[37] labeling_0.3     stringi_1.4.5    lazyeval_0.2.2   munsell_0.5.0   
[41] crayon_1.3.4    

[yutannihilation]

Curious. I saw several people reported this error about data.frame(), but haven't figured out how to reproduce. Thanks for the reprex.

[thomasp85]

There are a couple of things going on in your example that makes it hard to figure out if there are any problems in ggplot2 directly.

• loadRDS() -> readRDS()
• The pipe has precedence over plus so you are piping geom_bar() into saveRDS(), not the plot.
• I cannot reproduce your error message by fixing the above issues.

Can you try to fix your example up. run it from a clean session and confirm that the error message persists

[yutannihilation]

Ah, I too cannot reproduce the error... sorry

[yutannihilation]

@nick-youngblut Could you provide a reproducible example? Otherwise, we have no choice but to close this issue.

[nick-youngblut]

Sorry for the lack of reproducibility. It seems that the underlying cause was due to using saveRDS in one conda env running R and loading the saved object in a much different conda env running R (e.g., different R versions & package versions). I'm just going to try standardizing my R setups inside each conda env.

[yutannihilation]

e.g., different R versions & package versions

Then, it might be because you saved the plot created with some old version of ggplot2 where the internal use of data.frame() was not restricted (now it is). I recommend you to use renv to fix the versions.