CODEWITHSUNDEEP
rdplyr
AEF
AEF

Reputation: 5650

dplyr::mutate_if fails when column names contain spaces

I have the following situation: I have a tibble and a character vector that conatins some of said tibbles column names. Now I use dplyr::mutate_if to set those columns to 0, whose names are contained in the character vector. I do this by supplying a logical vector as predicate.

Up to now this has worked:

library(tidyverse)

test_OK <-
  tribble(~"Var1", ~"Var2", ~"Var3",
          11,      12,      13,
          21,      22,      23,
          31,      32,      33)

vars_to_change_OK <- c("Var1", "Var3")

## Everything works out
test_OK %>% mutate_if(names(.) %in% vars_to_change_OK, funs(. * 0))

But now I've come around a tibble with spaces and numbers in the column names and the procedure fails:

test_FAIL <-
  tribble(~"Var 1", ~"Var 2", ~"Var 3",
            11,      12,      13,
            21,      22,      23,
            31,      32,      33)

vars_to_change_FAIL <- c("Var 1", "Var 3")

## Error [...] unexpected numeric constant
test_FAIL %>% mutate_if(names(.) %in% vars_to_change_FAIL, funs(. * 0))

Note that the only difference are the column names which changed from Var[0-9] to Var [0-9].

I'd like to know if this is a bug in dplyr or if I have been using the mutate_if function in an unintended way. Furthermore I'm interested in how and why the error occurs. Thanks in advance for any insights!

PS.: I am aware that there is an easy base R workaround like this:

###############################
## workaround in base R:
test_FAIL[, match(vars_to_change_FAIL, names(test_FAIL))] <- 0
test_FAIL

I am however interested in why the above doesn't work in dplyr.

Update: I've added my sessionInfo() in case the version is important:

R version 3.4.0 (2017-04-21)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows >= 8 x64 (build 9200)

Matrix products: default

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

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

other attached packages:
 [1] dplyr_0.5.0          purrr_0.2.2          readr_1.1.0          tidyr_0.6.1          tibble_1.3.0        
 [6] ggplot2_2.2.1        tidyverse_1.1.1      dygraphs_1.1.1.4     shiny_1.0.3          RevoUtilsMath_10.0.0

loaded via a namespace (and not attached):
 [1] Rcpp_0.12.10     cellranger_1.1.0 compiler_3.4.0   plyr_1.8.4       forcats_0.2.0    tools_3.4.0      digest_0.6.12   
 [8] lubridate_1.6.0  jsonlite_1.4     nlme_3.1-131     gtable_0.2.0     lattice_0.20-35  psych_1.7.3.21   DBI_0.6-1       
[15] parallel_3.4.0   haven_1.0.0      xml2_1.1.1       httr_1.2.1       stringr_1.2.0    hms_0.3          RevoUtils_10.0.4
[22] htmlwidgets_0.8  grid_3.4.0       R6_2.2.0         readxl_1.0.0     foreign_0.8-67   modelr_0.1.0     reshape2_1.4.2  
[29] magrittr_1.5     scales_0.4.1     htmltools_0.3.6  rvest_0.3.2      assertthat_0.2.0 mnormt_1.5-5     colorspace_1.3-2
[36] mime_0.5         xtable_1.8-2     httpuv_1.3.3     stringi_1.1.5    lazyeval_0.2.0   munsell_0.4.3    broom_0.4.2     
[43] zoo_1.8-0

Upvotes: 2

Views: 1628

Answers (1)

akrun
akrun

Reputation: 887128

We can use mutate_at and place the 'vars_to_change_FAIL' inside vars

test_FAIL %>% 
      mutate_at(vars(vars_to_change_FAIL), funs(. * 0))
#  A tibble: 3 x 3
#   `Var 1` `Var 2` `Var 3`
#    <dbl>   <dbl>   <dbl>
#1       0      12       0
#2       0      22       0
#3       0      32       0

Upvotes: 4

Related Questions