Repository navigation
fails with ggh4x::facet_wrap2(trim_blank=FALSE) #442
Description
When using ggh4x::facet_wrap2 in a subplot and combining, it works only if trim_blank=TRUE (the default); if I try to preserve blanks then it presents a blank canvas.
library(ggplot2)
library(patchwork)
spldat <- data.frame(x=c(1, 1, 2), y = c(1, 2, 2)) |> split(~ x)
plots <- lapply(spldat, \(dat) ggplot(dat, aes(x, y)) + geom_point() +
ggh4x::facet_wrap2(y ~ ., ncol = 2, trim_blank = TRUE))
Reduce(`/`, plots)
plots <- lapply(spldat, \(dat) ggplot(dat, aes(x, y)) + geom_point() +
ggh4x::facet_wrap2(y ~ ., ncol = 2, trim_blank = FALSE))
Reduce(`/`, plots)(empty for me)
plots[[1]]
plots[[2]]
session info
xfun::session_info("patchwork")
# R version 4.4.3 (2025-02-28)
# Platform: aarch64-apple-darwin20
# Running under: macOS Sequoia 15.5
# Locale: en_US.UTF-8 / en_US.UTF-8 / en_US.UTF-8 / C / en_US.UTF-8 / en_US.UTF-8
# Package version:
# cli_3.6.5 farver_2.1.2 ggplot2_3.5.2 glue_1.8.0 graphics_4.4.3 grDevices_4.4.3 grid_4.4.3 gtable_0.3.6
# isoband_0.2.7 labeling_0.4.3 lattice_0.22.6 lifecycle_1.0.4 magrittr_2.0.3 MASS_7.3.64 Matrix_1.7.2 methods_4.4.3
# mgcv_1.9.1 nlme_3.1.167 patchwork_1.3.1 pillar_1.10.2 pkgconfig_2.0.3 R6_2.6.0 RColorBrewer_1.1.3 rlang_1.1.6
# scales_1.4.0 splines_4.4.3 stats_4.4.3 tibble_3.3.0 utf8_1.2.4 utils_4.4.3 vctrs_0.6.5 viridisLite_0.4.2
# withr_3.0.2
Activity
teunbrand commentedon Aug 11, 2025
It might be that one of the assumptions of patchwork is to have a grob present (even if it is a
zeroGrob()) at a panel position, whereas this is not the case withggh4x::facet_wrap2(trim_blank = FALSE). It might be more appropriate for ggh4x to replace the empty panels withzeroGrob()s than to remove the panels.r2evans commentedon Aug 11, 2025
I figured it'd be something special with the empty space.
r2evans commentedon Apr 8, 2026
Any traction here in patchwork or in ggh4x#202? @teunbrand
teunbrand commentedon Apr 8, 2026
I explored this problem a while ago but haven't found a satisfactory solution. Leaving in dummy grobs to satisfy the gtable assumptions would disrupt the current mechanism too much for me to commit to that.
r2evans commentedon Apr 9, 2026
For the record (future me + anybody else who may be interested, until something more "proper" may be implemented), an interim solution for me is to use
ggplot2::facet_wrap()for the normal plot, and add another similar plot with the correct number of panes withtheme_void()(completely empty). It's imperfect (the pane alignment is off a little in some of my plots though not below), but it's workable for now.