purrr::list_cbind() and vctrs::vec_cbind() returning a data.frame instead of a sf object

[elipousson]

Combining a simple feature object with a data.frame using purrr::list_cbind() or vctrs::vec_cbind() returns a data.frame instead of an sf object as expected.

nc <- sf::st_read(system.file("shape/nc.shp", package = "sf"))
#> Reading layer `nc' from data source 
#>   `/Users/elipousson/Library/R/arm64/4.2/library/sf/shape/nc.shp' 
#>   using driver `ESRI Shapefile'
#> Simple feature collection with 100 features and 14 fields
#> Geometry type: MULTIPOLYGON
#> Dimension:     XY
#> Bounding box:  xmin: -84.32385 ymin: 33.88199 xmax: -75.45698 ymax: 36.58965
#> Geodetic CRS:  NAD27

new_col <- data.frame("new_col" = seq_len(length(nc$NAME)))

nc_bind_cols <- dplyr::bind_cols(
  nc,
  new_col
)

class(nc_bind_cols)
#> [1] "sf"         "data.frame"

nc_list_cbind <- purrr::list_cbind(
  list(
    nc,
    new_col
  )
)

class(nc_list_cbind)
#> [1] "data.frame"

nc_vec_cbind <- vctrs::vec_cbind(
  nc,
  new_col
)

class(nc_vec_cbind)
#> [1] "data.frame"

^{Created on 2023-04-29 with reprex v2.0.2}

I'm pretty sure that both approaches worked with sf objects previously so I'm unsure if there was a change in vctrs or a change in sf.

R version 4.2.3 (2023-03-15)
Platform: aarch64-apple-darwin20 (64-bit)
Running under: macOS Big Sur 11.7.1

Matrix products: default
LAPACK: /Library/Frameworks/R.framework/Versions/4.2-arm64/Resources/lib/libRlapack.dylib

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

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

other attached packages:
[1] sinew_0.4.0        pak_0.4.0          devtools_2.4.5     usethis_2.1.6.9000

loaded via a namespace (and not attached):
 [1] tidyselect_1.2.0   remotes_2.4.2      rematch2_2.1.2     purrr_1.0.1       
 [5] sf_1.0-12          generics_0.1.3     vctrs_0.6.2        miniUI_0.1.1.1    
 [9] htmltools_0.5.5    yaml_2.3.7         utf8_1.2.3         rlang_1.1.0.9000  
[13] pkgbuild_1.4.0     e1071_1.7-13       urlchecker_1.0.1   later_1.3.0       
[17] pillar_1.9.0       glue_1.6.2         DBI_1.1.3          sessioninfo_1.2.2 
[21] lifecycle_1.0.3    stringr_1.5.0      htmlwidgets_1.6.2  memoise_2.0.1     
[25] callr_3.7.3        fastmap_1.1.1      httpuv_1.6.9       ps_1.7.4          
[29] class_7.3-21       fansi_1.0.4        Rcpp_1.0.10        KernSmooth_2.23-20
[33] xtable_1.8-4       promises_1.2.0.1   classInt_0.4-9     cachem_1.0.7      
[37] pkgload_1.3.2      mime_0.12          fs_1.6.1           sos_2.1-4         
[41] brew_1.0-8         digest_0.6.31      stringi_1.7.12     processx_3.8.0    
[45] dplyr_1.1.1        shiny_1.7.4        grid_4.2.3         cli_3.6.1         
[49] tools_4.2.3        magrittr_2.0.3     proxy_0.4-27       tibble_3.2.1      
[53] profvis_0.3.7      crayon_1.5.2       pkgconfig_2.0.3    ellipsis_0.3.2    
[57] prettyunits_1.1.1  rstudioapi_0.14    R6_2.5.1           units_0.8-1       
[61] compiler_4.2.3    

 GEOS           GDAL         proj.4 GDAL_with_GEOS     USE_PROJ_H           PROJ 
 "3.11.0"        "3.5.3"        "9.1.0"         "true"         "true"        "9.1.0" 

[francisbarton]

Hi @elipousson I have just seen this. I already raised a similar issue (though relating to list_rbind()) on the purrr repo. (Though a few days after you raised this issue!)
tidyverse/purrr#1076

[elipousson]

It wouldn't be a big issue but purrr::list_rbind() is being recommended as the replacement for the deprecated purrr::map_dfr() (and the row-binding functions have the same issue as the column-binding function) so I expect there may be some confused users. It is relatively straightforward to fix in a script but I'm not sure how it could be handled by sf or if vctrs or purrr would allow this edge case.

Here is an example showind the issue with purrr::list_rbind() (and the simple fix):

nc <- sf::read_sf(system.file("shape/nc.shp", package = "sf"))

nc_bind_rows <- dplyr::bind_rows(
  nc[2, ],
  nc[1, ]
)

class(nc_bind_rows)
#> [1] "sf"         "tbl_df"     "tbl"        "data.frame"

nc_vec_rbind <- vctrs::vec_rbind(
  nc[2, ],
  nc[1, ]
)

class(nc_vec_rbind)
#> [1] "data.frame"

class(sf::st_as_sf(nc_vec_rbind))
#> [1] "sf"         "data.frame"

^{Created on 2023-06-26 with reprex v2.0.2}

[francisbarton]

I do think this is an issue with vctrs or purrr not with sf though.

[hadley]

To be clear, map_dfr() is not deprecated; it's superseded.

[elipousson]

Thanks for that reminder!