`'character' chr` is not quite the same as `chr`

[courtiol]

sf objects seem to have their own version of "character" ('character' chr rather than chr).
Depending on how one extracts a column, the 'character' chr stays in and it can create issues with other R functions.
At least it does so with vctrs and thus dplyr as shown below.

I did not check if drop is a generic which a sf method could fine tune, or if it could create issues beyond vctrs.

Strangely base R seems to interpret 'character' chr as a genuine chr.

library(sf)
#> Linking to GEOS 3.14.1, GDAL 3.11.5, PROJ 9.6.2; sf_use_s2() is TRUE
library(dplyr, warn.conflicts = FALSE)

pt1 = st_point(c(0,1))
pt2 = st_point(c(1,1))
d = data.frame(a = letters[1:2])
d$geom = st_sfc(pt1, pt2)
df = st_as_sf(d)

str(df[["a"]]) # same as str(st_drop_geometry(df)[, "a"])
#>  chr [1:2] "a" "b"
str(df[, "a", drop = TRUE])
#>  'character' chr [1:2] "a" "b"

class(df[["a"]])
#> [1] "character"
class(df[, "a", drop = TRUE])
#> [1] "character"
is.character(df[["a"]])
#> [1] TRUE
is.character(df[, "a", drop = TRUE])
#> [1] TRUE

full_join(data.frame(foo = df[["a"]]),
          data.frame(foo = df[["a"]]))
#> Joining with `by = join_by(foo)`
#>   foo
#> 1   a
#> 2   b

full_join(data.frame(foo = df[, "a", drop = TRUE]),
          data.frame(foo = df[["a"]]))
#> Joining with `by = join_by(foo)`
#> Error in `stop_native_implementation()`:
#> ! `vec_ptype2.character.character()` is implemented at C level.
#>   This R function is purely indicative and should never be called.
#> ℹ This is an internal error that was detected in the vctrs package.
#>   Please report it at <https://github.com/r-lib/vctrs/issues> with a reprex
#>   (<https://tidyverse.org/help/>) and the full backtrace.

sessionInfo()
#> R version 4.5.2 (2025-10-31)
#> Platform: x86_64-redhat-linux-gnu
#> Running under: Fedora Linux 43 (KDE Plasma Desktop Edition)
#> 
#> Matrix products: default
#> BLAS/LAPACK: FlexiBLAS OPENBLAS-OPENMP;  LAPACK version 3.12.1
#> 
#> locale:
#>  [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
#>  [3] LC_TIME=en_GB.UTF-8        LC_COLLATE=en_US.UTF-8    
#>  [5] LC_MONETARY=en_GB.UTF-8    LC_MESSAGES=en_US.UTF-8   
#>  [7] LC_PAPER=en_GB.UTF-8       LC_NAME=C                 
#>  [9] LC_ADDRESS=C               LC_TELEPHONE=C            
#> [11] LC_MEASUREMENT=en_GB.UTF-8 LC_IDENTIFICATION=C       
#> 
#> time zone: Europe/Berlin
#> tzcode source: system (glibc)
#> 
#> attached base packages:
#> [1] stats     graphics  grDevices utils     datasets  methods   base     
#> 
#> other attached packages:
#> [1] dplyr_1.1.4 sf_1.0-23  
#> 
#> loaded via a namespace (and not attached):
#>  [1] vctrs_0.6.5        cli_3.6.5          knitr_1.50         rlang_1.1.6       
#>  [5] xfun_0.54          DBI_1.2.3          KernSmooth_2.23-26 generics_0.1.4    
#>  [9] glue_1.8.0         htmltools_0.5.8.1  e1071_1.7-16       rmarkdown_2.30    
#> [13] grid_4.5.2         tibble_3.3.0       evaluate_1.0.5     classInt_0.4-11   
#> [17] fastmap_1.2.0      yaml_2.3.10        lifecycle_1.0.4    compiler_4.5.2    
#> [21] fs_1.6.6           pkgconfig_2.0.3    Rcpp_1.1.0         rstudioapi_0.17.1 
#> [25] digest_0.6.39      R6_2.6.1           tidyselect_1.2.1   class_7.3-23      
#> [29] reprex_2.1.1       pillar_1.11.1      magrittr_2.0.4     tools_4.5.2       
#> [33] withr_3.0.2        proxy_0.4-27       units_1.0-0

^{Created on 2025-11-29 with reprex v2.1.1}

[edzer]

We see

> attr(df[,"a",drop=TRUE], "class")
[1] "character"

which definitely seems unintended.

[edzer]

Should be fixed now; thanks!