Error in if (is.na(x)) { : argument is of length zero

Prework

[x] Read and abide by drake's code of conduct.
[x] Search for duplicates among the existing issues, both open and closed.
[x] If you think your issue has a quick and definite solution, consider posting to Stack Overflow under the drake-r-package tag. (If you anticipate extended follow-up and discussion, you are already in the right place!)

Description

When I run vis_drake_graph() ormake(), I get the error: Error in if (is.na(x)) { : argument is of length zero. Please note downgrading to drake 7.8.0 resolves the issue. This only happens with the latest dev version.

Reproducible example

This is not straightforward because the code is sensitive to my organization. I have tried to replicate the issue without success - although I did accidentally replicate another issue #1086!

I will try again soon.

Desired result

Typical behaviour for vis_drake_graph() and make()

Session info

R version 3.5.2 (2018-12-20)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 17134)

Matrix products: default

locale:
[1] LC_COLLATE=English_Canada.1252  LC_CTYPE=English_Canada.1252    LC_MONETARY=English_Canada.1252
[4] LC_NUMERIC=C                    LC_TIME=English_Canada.1252    

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

other attached packages:
 [1] tidyselect_0.2.5    here_0.1            forcats_0.4.0       stringr_1.4.0       dplyr_0.8.3        
 [6] purrr_0.3.2         readr_1.3.1         tidyr_1.0.0         tibble_2.1.3        tidyverse_1.2.1    
[11] pROC_1.15.3         caret_6.0-84        ggplot2_3.2.1       lattice_0.20-38     broom_0.5.2        
[16] modelr_0.1.5        feather_0.3.5       digest_0.6.23       e1071_1.7-2         randomForest_4.6-14
[21] RWeka_0.4-41        rpart_4.1-13        gbm_2.1.5           fs_1.3.1            lubridate_1.7.4    
[26] janitor_1.2.0       rvest_0.3.4         xml2_1.2.2          httr_1.4.1          readxl_1.3.1       
[31] haven_2.1.1         fuzzyjoin_0.1.5     glue_1.3.1          ROracle_1.3-1       DBI_1.0.0          
[36] drake_7.8.0.9000   

loaded via a namespace (and not attached):
 [1] colorspace_1.4-1     ellipsis_0.3.0       class_7.3-14         rprojroot_1.3-2      rstudioapi_0.10     
 [6] remotes_2.1.0        prodlim_2018.04.18   codetools_0.2-15     splines_3.5.2        pkgload_1.0.2       
[11] zeallot_0.1.0        jsonlite_1.6         rJava_0.9-11         dbplyr_1.4.2         compiler_3.5.2      
[16] keyring_1.1.0        backports_1.1.4      assertthat_0.2.1     Matrix_1.2-15        lazyeval_0.2.2      
[21] cli_1.1.0            htmltools_0.4.0      visNetwork_2.0.8     prettyunits_1.0.2    tools_3.5.2         
[26] igraph_1.2.4.1       gtable_0.3.0         reshape2_1.4.3       Rcpp_1.0.2           cellranger_1.1.0    
[31] vctrs_0.2.0          nlme_3.1-137         iterators_1.0.12     timeDate_3043.102    dbmanager_0.0.0.9001
[36] gower_0.2.1          ps_1.3.0             testthat_2.2.1       lifecycle_0.1.0      RWekajars_3.9.3-2   
[41] devtools_2.2.1       MASS_7.3-51.1        scales_1.0.0         ipred_0.9-9          hms_0.5.1           
[46] yaml_2.2.0           memoise_1.1.0        gridExtra_2.3        stringi_1.4.3        desc_1.2.0          
[51] foreach_1.4.7        filelock_1.0.2       pkgbuild_1.0.6       lava_1.6.6           storr_1.2.1         
[56] rlang_0.4.0          pkgconfig_2.0.2      htmlwidgets_1.5      recipes_0.1.7        processx_3.4.1      
[61] plyr_1.8.4           magrittr_1.5         R6_2.4.0             generics_0.0.2       base64url_1.4       
[66] txtq_0.1.6           pillar_1.4.2         withr_2.1.2          survival_2.43-3      nnet_7.3-12         
[71] crayon_1.3.4         usethis_1.5.1        grid_3.5.2           data.table_1.12.2    callr_3.3.1         
[76] ModelMetrics_1.2.2   stats4_3.5.2         munsell_0.5.0        sessioninfo_1.1.1

At first glance, this seemed like a cryptic bug I would need a reprex for. But the only occurrence of "if (is.na(x)) {" is here:

https://github.com/ropensci/drake/blob/a12283917fdce1f706c88bff80e55d9bcf70eaf9/R/utils.R#L295

Changing is.na() to anyNA() should be safer without much loss in speed.

library(microbenchmark)
microbenchmark(
  a = anyNA("x"),
  b = is.na("x")
)
#> Unit: nanoseconds
#>  expr min lq   mean median uq  max neval
#>     a  52 58 105.37   65.0 68 4010   100
#>     b  66 71  99.09   78.5 83 1913   100
system.time(replicate(1e5, anyNA(sample.int(1e5, 1))))
#>    user  system elapsed 
#>   0.283   0.011   0.294
system.time(replicate(1e5, is.na(sample.int(1e5, 1))))
#>    user  system elapsed 
#>   0.286   0.000   0.286

^{Created on 2019-12-09 by the reprex package (v0.3.0)}

I will make the change. If it solves the problem, please let me know. If not, please post a traceback(). A reprex is the best thing, but probably difficult in your situation.

ropensci / drake