Allow missing tag level

[jeanettelt]

Hi there,

My project team and I have found this package much more useful and intuitive than cowplot, so thank you!

We've been using patchwork to combine multiple plots into one figure. Our largest figure is a 3x3 grid where each plot is a combined 2x1 plot, so there's actually a total of 18 ggplot objects.

We were hoping to use the plot_layout(tag_level = 'new') feature to skip labeling our subplots, but the skipped subplots also skip the associated label letters. Here is a simplified example:

library(tidyverse)
library(patchwork)
p1 <- ggplot(mtcars) + geom_point(aes(mpg, disp)) + ggtitle("Main plot 1")
p2 <- ggplot(mtcars) + geom_boxplot(aes(gear, disp, group = gear)) + ggtitle("Main plot 2")
h1 <- ggplot(mtcars) + geom_histogram(aes(mpg)) + plot_layout(tag_level = 'new') + ggtitle("Subplot 1")
h2 <- ggplot(mtcars) + geom_histogram(aes(gear)) + scale_x_continuous(limits = c(2.5, 5.5), expand = c(0, 0)) + ggtitle("Subplot 2") + plot_layout(tag_level = 'new')
((p1 / h1) | (p2 / h2)) + plot_annotation(tag_levels = c('A'))
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#> Warning: Removed 2 rows containing missing values (geom_bar).

^{Created on 2020-01-17 by the reprex package (v0.3.0)}

We've instead used ggplot2's labs(title = ) feature, like in the modified example below, to manually create labels for our nine plots, but were hoping for a simpler solution through patchwork.

library(tidyverse)
library(patchwork)
p1 <- ggplot(mtcars) + geom_point(aes(mpg, disp)) + ggtitle("Main plot 1")
p2 <- ggplot(mtcars) + geom_boxplot(aes(gear, disp, group = gear)) + ggtitle("Main plot 2")
h1 <- ggplot(mtcars) + geom_histogram(aes(mpg)) + plot_layout(tag_level = 'new') + ggtitle("Subplot 1")
h2 <- ggplot(mtcars) + geom_histogram(aes(gear)) + scale_x_continuous(limits = c(2.5, 5.5), expand = c(0, 0)) + ggtitle("Subplot 2")
(((p1 + labs(title = "A")) / h1) | ((p2 + labs(title = "B")) / h2))
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#> Warning: Removed 2 rows containing missing values (geom_bar).

^{Created on 2020-01-17 by the reprex package (v0.3.0)}

[kendonB]

@thomasp85 this fix fails in the special case where the user arranges the plots using the + operator for the first plot:

library(tidyverse)
library(patchwork)
p1 <- ggplot(mtcars) + geom_point(aes(mpg, disp)) + ggtitle("Main plot 1")
p2 <- ggplot(mtcars) + geom_boxplot(aes(gear, disp, group = gear)) + ggtitle("Main plot 2")
h1 <- ggplot(mtcars) + geom_histogram(aes(mpg)) + plot_layout(tag_level = 'new') + ggtitle("Subplot 1")
h2 <- ggplot(mtcars) + geom_histogram(aes(gear)) + scale_x_continuous(limits = c(2.5, 5.5), expand = c(0, 0)) + ggtitle("Subplot 2") + plot_layout(tag_level = 'new')

h1 + h2 + p1 + p2 + plot_annotation(tag_levels = c('A'))
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#> Warning: Removed 2 rows containing missing values (geom_bar).

^{Created on 2021-03-03 by the reprex package (v1.0.0)}

[kendonB]

A bit of a hack is the following solution using the text-based layout feature:

library(tidyverse)
library(patchwork)
p1 <- ggplot(mtcars) + geom_point(aes(mpg, disp)) + ggtitle("Main plot 1")
p2 <- ggplot(mtcars) + geom_boxplot(aes(gear, disp, group = gear)) + ggtitle("Main plot 2")
h1 <- ggplot(mtcars) + geom_histogram(aes(mpg)) + plot_layout(tag_level = 'new') + ggtitle("Subplot 1")
h2 <- ggplot(mtcars) + geom_histogram(aes(gear)) + scale_x_continuous(limits = c(2.5, 5.5), expand = c(0, 0)) + ggtitle("Subplot 2") + plot_layout(tag_level = 'new')
layout = "
          CD
          AB
         "
p1 + p2 + h1 + h2 + plot_annotation(tag_levels = c('A')) + plot_layout(design = layout)
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
#> Warning: Removed 2 rows containing missing values (geom_bar).

^{Created on 2021-03-03 by the reprex package (v1.0.0)}

[kendonB]

@thomasp85 I landed here from Google, just adding the syntax that took me a while to figure out:

plot_annotation(tag_levels = list(c("A", "B", "C", "D", "E")))

This, for example, deletes the F if it were there

[iamakhilverma]

@thomasp85 I landed here from Google, just adding the syntax that took me a while to figure out:

plot_annotation(tag_levels = list(c("A", "B", "C", "D", "E")))

This, for example, deletes the F if it were there

Good hack, Thanks a lot!
For me ignore_tag = T wasn't working as intended, so, used this hack to modify my annotation to exclude the insets-

plot_annotation(tag_levels = list(c("A", "", "B", "", "C", "", "D", "", "E", "", "F", "", "G", "H")))