ggpattern doesn't seem to work with latest DEV version

[davidhodge931]

library(palmerpenguins)
library(tidyverse)
library(ggpattern)

penguins |>
  group_by(species, sex) |>
  summarise(across(body_mass_g, \(x) mean(x))) |>
  ggplot(    mapping = aes(y = species,
                           x = body_mass_g,
                           fill = sex, 
                           pattern = sex)) +
  geom_col_pattern()
#> `summarise()` has grouped output by 'species'. You can override using the
#> `.groups` argument.
#> Warning: The `scale_name` argument of `discrete_scale()` is deprecated as of ggplot2
#> 3.5.0.
#> This warning is displayed once every 8 hours.
#> Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
#> generated.
#> Warning: Removed 2 rows containing missing values or values outside the scale range
#> (`geom_col_pattern()`).
#> Error in if (pat == "regular_polygon" && is.numeric(params$pattern_shape)) params$pattern_shape <- "convex6": missing value where TRUE/FALSE needed

^{Created on 2023-12-22 with reprex v2.0.2}

[teunbrand]

This is a more minimal reprex

devtools::load_all("~/packages/ggplot2")
# or:
# library(ggplot2) # if dev version is installed
#> ℹ Loading ggplot2
library(ggpattern)

df <- data.frame(x = c(1, 2, 3), y = c(1, 2, NA), fill = c("A", "B", NA))

ggplot(df, aes(x, y, fill = fill, pattern = fill)) +
  geom_col_pattern()
#> Warning: The `scale_name` argument of `discrete_scale()` is deprecated as of ggplot2
#> 3.5.0.
#> This warning is displayed once every 8 hours.
#> Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
#> generated.
#> Warning: Removed 1 row containing missing values or values outside the scale range
#> (`geom_col_pattern()`).
#> Error in if (pat == "regular_polygon" && is.numeric(params$pattern_shape)) params$pattern_shape <- "convex6": missing value where TRUE/FALSE needed

Seems to be solved by setting a na.value in the scale:

last_plot() + scale_pattern_discrete(na.value = "crosshatch")
#> Warning: Removed 1 row containing missing values or values outside the scale range
#> (`geom_col_pattern()`).

^{Created on 2024-01-02 with reprex v2.0.2}

[teunbrand]

Seems to be caused by #5519, where missing values are no longer silently removed by position_stack().
You can trigger the same error (in a different location) with CRAN ggplot2 by setting position = "identity", which also doesn't remove NA values.

library(ggplot2)
library(ggpattern)

df <- data.frame(x = c(1, 2, 3), y = c(1, 2, NA), fill = c("A", "B", NA))

ggplot(df, aes(x, y, fill = fill, pattern = fill)) +
  geom_col_pattern(position = 'identity')
#> Warning: Removed 1 rows containing missing values (`geom_col_pattern()`).
#> Error in if (pat == "regular_polygon" && is.numeric(params$pattern_shape)) params$pattern_shape <- "convex6": missing value where TRUE/FALSE needed

^{Created on 2024-01-02 with reprex v2.0.2}

In any case, as we're unlikely to revert #5519, it seems like the most suitable fix is for {ggpattern} to set a na.value for the discrete scale.