Sizing geom_sf point data with variable containing NAs

[daniel-barnett]

Trying to plot point data with geom_sf() that has a mapping to a variable that contains NAs creates an error. This occurs both in the CRAN 3.2.1 and development versions.

library(ggplot2)
packageVersion("ggplot2")
#> [1] '3.2.1.9000'

# From geom_sf() examples in docs
nc <- sf::st_read(system.file("shape/nc.shp", package = "sf"), quiet = TRUE)
nc_3857 <- sf::st_transform(nc, "+init=epsg:3857")

nc_3857$mid <- sf::st_centroid(nc_3857$geometry)

ggplot(nc_3857) +
  geom_sf(aes(geometry = mid, size = AREA))

nc_3857$AREA[10] <- NA

ggplot(nc_3857) +
  geom_sf(aes(geometry = mid, size = AREA))
#> Error in checkNA("fontsize"): mixture of missing and non-missing values for fontsize

sessionInfo()
#> R version 3.6.1 (2019-07-05)
#> Platform: x86_64-w64-mingw32/x64 (64-bit)
#> Running under: Windows 7 x64 (build 7601) Service Pack 1
#> 
#> Matrix products: default
#> 
#> locale:
#> [1] LC_COLLATE=English_New Zealand.1252 
#> [2] LC_CTYPE=English_New Zealand.1252   
#> [3] LC_MONETARY=English_New Zealand.1252
#> [4] LC_NUMERIC=C                        
#> [5] LC_TIME=English_New Zealand.1252    
#> 
#> attached base packages:
#> [1] stats     graphics  grDevices utils     datasets  methods   base     
#> 
#> other attached packages:
#> [1] ggplot2_3.2.1.9000
#> 
#> loaded via a namespace (and not attached):
#>  [1] Rcpp_1.0.2         pillar_1.4.2       compiler_3.6.1    
#>  [4] highr_0.8          class_7.3-15       tools_3.6.1       
#>  [7] digest_0.6.20      evaluate_0.14      tibble_2.1.3      
#> [10] gtable_0.3.0       pkgconfig_2.0.2    rlang_0.4.0       
#> [13] DBI_1.0.0          yaml_2.2.0         xfun_0.8          
#> [16] e1071_1.7-2        withr_2.1.2        dplyr_0.8.3       
#> [19] stringr_1.4.0      knitr_1.23         classInt_0.4-1    
#> [22] grid_3.6.1         tidyselect_0.2.5   glue_1.3.1        
#> [25] sf_0.7-7           R6_2.4.0           rmarkdown_1.14    
#> [28] purrr_0.3.2        magrittr_1.5       units_0.6-3       
#> [31] scales_1.0.0       htmltools_0.3.6    assertthat_0.2.1  
#> [34] colorspace_1.4-1   labeling_0.3       KernSmooth_2.23-15
#> [37] stringi_1.4.3      lazyeval_0.2.2     munsell_0.5.0     
#> [40] crayon_1.3.4

^{Created on 2019-08-13 by the reprex package (v0.3.0)}

[yutannihilation]

Hmm, this change (I'm not sure if we call this a "regression") seems introduced in v3.2.0... The code used to work before:

library(ggplot2)
#> Registered S3 methods overwritten by 'ggplot2':
#>   method         from 
#>   [.quosures     rlang
#>   c.quosures     rlang
#>   print.quosures rlang
packageVersion("ggplot2")
#> [1] '3.1.1'

nc <- sf::st_read(system.file("shape/nc.shp", package = "sf"), quiet = TRUE)
nc_3857 <- sf::st_transform(nc, "+init=epsg:3857")

nc_3857$mid <- sf::st_centroid(nc_3857$geometry)

nc_3857$AREA[10] <- NA

ggplot(nc_3857) +
  geom_sf(aes(geometry = mid, size = AREA))

^{Created on 2019-08-13 by the reprex package (v0.3.0)}

The error comes from here in sf_grob():

ggplot2/R/geom-sf.R

Lines 159 to 162 in fc8f7d5

	gp <- gpar(
	col = col, fill = fill, fontsize = fontsize, lwd = lwd, lty = lty,
	lineend = lineend, linejoin = linejoin, linemitre = linemitre
	)

Since grid::gpar() accepts a single NA, containing NAs hadn't become a problem until sf_grob() got vectorized by #3164.

grid::gpar(fontsize = NA)
#> named list()
grid::gpar(fontsize = c(1, NA))
#> Error in checkNA("fontsize"): mixture of missing and non-missing values for fontsize

^{Created on 2019-08-13 by the reprex package (v0.3.0)}

Now that it's vectorized, it seems we need to replace NAs on ggplot2's side to let it be drawn. But, I'm not sure if NA-size points should be drawn or not. Maybe it should not, similarly to #2757 (comment)?

[yutannihilation]

But, I'm not sure if NA-size points should be drawn or not.

Now I'm sure it should not. Probably, GeomSf needs to have its own handle_na() because non_missing_aes varies, depending the type of the geometry.

ggplot2/R/geom-.r

Lines 67 to 72 in b842024

	handle_na = function(self, data, params) {
	remove_missing(data, params$na.rm,
	c(self$required_aes, self$non_missing_aes),
	snake_class(self)
	)
	},

[yutannihilation]

GeomSf needs to have its own handle_na()

Sorry, I was wrong here. It seems GeomSf should always have non_missing_aes() as sf_grob() always passes fontsize.

[lock]

This old issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with reprex) and link to this issue. https://reprex.tidyverse.org/