drop.units = FALSE is not working well for points

[Nowosad]

@mtennekes see my reprex:

library(tmap)

# works
tm_shape(NLD_muni) +
  tm_borders(lwd = 3) +
  tm_facets(by = "province", drop.units = FALSE)

# works
tm_shape(NLD_muni) +
  tm_dots()

# works (for symbols), but wrongly
tm_shape(NLD_muni) +
  tm_symbols() +
  tm_facets(by = "province", drop.units = FALSE)

# fails (for dots)
tm_shape(NLD_muni) +
  tm_dots() +
  tm_facets(by = "province", drop.units = FALSE)
#> Error in `[.data.table`(res, , `:=`(names(defs), Map(replace, .SD, lapply(.SD, : .SDcols missing at the following indices: [3]

^{Created on 2025-11-28 with reprex v2.1.1}

[mtennekes]

@Nowosad The tm_dots error has been fixed. But what do you expect to see for

tm_shape(NLD_muni) +
  tm_symbols() +
  tm_facets(by = "province", drop.units = FALSE)

?

For polygons, the units that are not highlighted are not dropped, but grayed out. More precisely, value.null is applied here.

tmap_options("value.null")
$value.null
$value.null$fill
[1] "grey95"

$value.null$col
[1] "grey95"

$value.null$col.polygons
[1] "grey40"

$value.null$bgcol
[1] "grey95"

$value.null$lty
[1] "solid"

$value.null$lwd
[1] 0.2

$value.null$shape
[1] 20

$value.null$text
[1] ""

$value.null$fontface
NULL

$value.null$fill_alpha
[1] 1

$value.null$col_alpha
[1] 1

$value.null$bgcol_alpha
[1] 1

$value.null$size
[1] 0.2

$value.null$angle
[1] 0

$value.null$num
[1] 0

As you can see, the current 'out-of-selection' properties for symbols (and dots) are size = 0.2, and col/fill = gray.

Is this okay, or would you like to see something else as default?

[Nowosad]

Hi @mtennekes -- all the best for the New Year!

My expectation is that the polygons and dots examples will behave in the same way: for polygons, the out-of-selection polygons are still visible. I would like the same for dots/symbols: the highlighted points should be highlighted, but the rest of the points should still be visible.

[mtennekes]

It's not clear what sizes they should have. In your example we could (and probably should) take the same sizes as the constant values:

vnull = tmap_options("value.null")$value.null
vnull$size = 1
vnull$size.bubbles = 1.3333
vnull$size.squares = 1.3333
vnull$size.dots = .3
vnull$shape.squares = 22
vnull$shape.dots = 19
vnull$shape = 21

tmap_options(value.null = vnull)

tm_shape(NLD_muni) +
	tm_dots() +
	tm_facets(by = "province", drop.units = FALSE)

tm_shape(NLD_muni) +
	tm_symbols() +
	tm_facets(by = "province", drop.units = FALSE)

please test.

But how about facets where the size is a data variable?

[Nowosad]

But how about facets where the size is a data variable?

Good question. Should it not be the same (variable) as the highlighted data (but less visible)?