Specifying correct raster labels

[Nowosad]

Often, when you crop existing categorical raster data, the new object does not contain all of the original categories. This could be a problem when plotting the raster using some custom palettes and labels. Custom palettes still could be used in this situation by setting the raster values as names of the palette vector. However, the same operation does not work for labels. You can see some examples below, for data with values between 1 and 7 with 5 missing.

library(stars)
library(tmap)
tf = tempfile(fileext = ".zip")
td = tempdir()
download.file("https://github.com/r-tmap/tmap/files/7380213/tmc.zip", tf)
unzip(tf, exdir = td)
lc = read_stars(paste0(td, "/lc.tif"))
lc_palette_df = read.csv(paste0(td, "/legend.csv"))

# wrong colors for 6 and 7
tm_shape(lc) +
  tm_raster(style = "cat", palette = lc_palette_df$color)

# correct colors
my_color = lc_palette_df$color
names(my_color) = lc_palette_df$value
tm_shape(lc) +
  tm_raster(style = "cat", palette = my_color)

# wrong labels for 6 and 7
tm_shape(lc) +
  tm_raster(style = "cat", palette = lc_palette_df$color,
            labels = lc_palette_df$label)

# the fix for colors does not work for labels
my_label = lc_palette_df$label
names(my_label) = lc_palette_df$value
tm_shape(lc) +
  tm_raster(style = "cat", palette = my_color,
            labels = my_label)

tmc.zip

CC: @blakos1

[mtennekes]

It should work now. Have to check if this occurs in tmap4 as well.

[Nowosad]

Thanks!

[Nowosad]

@blakos1