Error in .subset(x, j) : invalid subscript type 'language' when using mutate_all

[adamryczkowski]

library(data.table)
library(dtplyr)
library(dplyr)
dt <- data.table(a=1:10, b=1:10, c=1:10)
dt %>%  mutate_all(as.integer)
# Error in .subset(x, j) : invalid subscript type 'language'

When we remove data.table nature of the dt, everything works:

library(data.table)
library(dtplyr)
library(dplyr)
dt <- data.table(a=1:10, b=1:10, c=1:10)
df <- as.data.frame(dt)
df %>% mutate_all(as.integer)

#     a  b  c
# 1   1  1  1
# 2   2  2  2
# 3   3  3  3
# 4   4  4  4
# 5   5  5  5
# 6   6  6  6
# 7   7  7  7
# 8   8  8  8
# 9   9  9  9
# 10 10 10 10

> sessionInfo()
R version 3.4.0 (2017-04-21)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 16.04.2 LTS

Matrix products: default
BLAS: /usr/lib/libblas/libblas.so.3.6.0
LAPACK: /usr/lib/lapack/liblapack.so.3.6.0

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

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

other attached packages:
[1] bindrcpp_0.1      data.table_1.10.4 dplyr_0.7.0       dtplyr_0.0.2.9000

loaded via a namespace (and not attached):
 [1] Rcpp_0.12.11     digest_0.6.12    withr_1.0.2      assertthat_0.2.0 R6_2.2.1         git2r_0.18.0     magrittr_1.5     httr_1.2.1       rlang_0.1.1     
[10] lazyeval_0.2.0   curl_2.6         devtools_1.13.2  tools_3.4.0      glue_1.1.0       compiler_3.4.0   memoise_1.1.0    bindr_0.1        tibble_1.3.3    

[MarkusBonsch]

Same for transmute with dplyr 0.7.1 under R 3.3.3:

library(data.table)
library(dplyr)
library(dtplyr)

test <- data.table(a=1, b=2)
transmute(test, c = 1)
# Error in .subset(x, j) : invalid subscript type 'language'

> sessionInfo()
R version 3.3.3 (2017-03-06)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 7 x64 (build 7601) Service Pack 1

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] dtplyr_0.0.2         dplyr_0.7.1          data.table_1.10.4    RevoUtilsMath_10.0.0 RevoUtils_10.0.3    
 [6] RevoMods_11.0.0      MicrosoftML_1.3.0    mrsdeploy_1.1.0      RevoScaleR_9.1.0     lattice_0.20-34     
[11] rpart_4.1-10        

loaded via a namespace (and not attached):
 [1] Rcpp_0.12.11           codetools_0.2-15       CompatibilityAPI_1.1.0 assertthat_0.1         foreach_1.4.3         
 [6] grid_3.3.3             R6_2.2.0               jsonlite_1.3           magrittr_1.5           rlang_0.1.1           
[11] lazyeval_0.2.0         curl_2.3               bindrcpp_0.2           iterators_1.0.8        tools_3.3.3           
[16] glue_1.1.1             pkgconfig_2.0.1        bindr_0.1              tibble_1.3.3           mrupdate_1.0.1     

[jreades]

I've also encountered this (as part of a long pipeline, and my example wouldn't meaningfully add to this) and the workaround posted above also worked (as.data.frame(dt)).

> sessionInfo()
R version 3.3.3 (2017-03-06)
Platform: x86_64-apple-darwin13.4.0 (64-bit)
Running under: macOS Sierra 10.12.5

locale:
[1] en_GB.UTF-8/en_GB.UTF-8/en_GB.UTF-8/C/en_GB.UTF-8/en_GB.UTF-8

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

other attached packages:
 [1] Hmisc_4.0-3     ggplot2_2.2.1   Formula_1.2-1   survival_2.40-1 lattice_0.20-34 zoo_1.8-0       dtplyr_0.0.2    bindrcpp_0.2    sf_0.5-2        reshape2_1.4.2 
[11] dplyr_0.7.1     plyr_1.8.4     

loaded via a namespace (and not attached):
 [1] Rcpp_0.12.11        RColorBrewer_1.1-2  bindr_0.1           base64enc_0.1-3     tools_3.3.3         digest_0.6.9        rpart_4.1-10        checkmate_1.8.2    
 [9] htmlTable_1.9       tibble_1.3.3        gtable_0.2.0        pkgconfig_2.0.1     rlang_0.1.1         Matrix_1.2-8        DBI_0.7             gridExtra_2.2.1    
[17] knitr_1.13          cluster_2.0.5       stringr_1.0.0       htmlwidgets_0.8     nnet_7.3-12         grid_3.3.3          glue_1.1.1          data.table_1.9.6   
[25] R6_2.1.2            foreign_0.8-67      latticeExtra_0.6-28 udunits2_0.13       magrittr_1.5        htmltools_0.3.6     backports_1.1.0     scales_0.4.1       
[33] units_0.4-5         splines_3.3.3       assertthat_0.1      colorspace_1.2-6    stringi_1.1.1       acepack_1.4.1       lazyeval_0.2.0      munsell_0.4.3      
[41] chron_2.3-47 

[Feakster]

The same problem is present for mutate_at():

Error in .subset(x, j) : invalid subscript type 'language'

[BabaYara]

One work around is to wrap your object in "as_tibble()" function to drop the data.table class attributes from the object but retain the benefits of a tibble object.

[MarkusBonsch]

What you essentially say is: don't use data.table if working with dplyr.
I personally will rather stop using dplyr if this problem doesn't get any attention.

[lionel-]

We may try to get an update during the summer, I'll talk to Hadley about dtplyr.

[MarkusBonsch]

Thank you.
It would also be good to have a statement about the future strategy:
is it envisaged to maintain data.table support for dplyr through dtplyr?
At the moment, this doesn't seem to be the case.

[otoomet]

A similar issue:

> a <- data.table(a=1:2, b=c("a", "b"))
> dplyr::group_by_(a, "b") %>% summarize(m=mean(a))
Error in .subset(x, j) : invalid subscript type 'language'

if wrapping a into as_tibble(), or if using NSE, it works as expected.

> packageVersion("dplyr")
[1] ‘0.7.3’
> packageVersion("dtplyr")
[1] ‘0.0.2’