使用R的patchwork组合多级绘图时面板间距不一致问题

使用R的patchwork包组合多张ggplot绘图时，出现面板错位及间距不一致的问题：第一组绘图（pA）的面板间距为双倍，其他组（pB、pC）为单倍，导致布局视觉不协调。代码及问题示意图如下：

library(ggplot2)
library(patchwork)

# Example data
mpg

# First set of plots
p1 <- ggplot(mpg, aes(class, cyl, group = manufacturer)) +
  geom_bar(stat = "identity", position = position_dodge())
p2 <- ggplot(mpg, aes(class, cyl, group = manufacturer)) +
  geom_line(aes(linetype = manufacturer, color = manufacturer), position = position_dodge2(width = 0.2), linewidth = 0.5)

pA <- p1 | p2
pA <- pA + plot_layout(axis_titles = 'collect')

# Second set of plots
p3 <- ggplot(mpg, aes(class, displ, group = manufacturer)) +
  geom_bar(stat = "identity", position = position_dodge())
p4 <- ggplot(mpg, aes(class, displ, group = manufacturer)) +
  geom_line(aes(linetype = manufacturer, color = manufacturer), position = position_dodge2(width = 0.2), linewidth = 0.5)

pB <- p3 | p4
pB <- pB + plot_layout(axis_titles = 'collect')

# Third set of plots
p5 <- ggplot(mpg, aes(class, year, group = manufacturer)) +
  geom_bar(stat = "identity", position = position_dodge())
p6 <- ggplot(mpg, aes(class, year, group = manufacturer)) +
  geom_line(aes(linetype = manufacturer, color = manufacturer), position = position_dodge2(width = 0.2), linewidth = 0.5)

pC <- p5 | p6
pC <- pC + plot_layout(axis_titles = 'collect')

# Combined final plot
pfail <- pA | pB | pC

pfail[[1]] <- pfail[[1]] + plot_layout(tag_level = 'new')
pfail[[2]] <- pfail[[2]] + plot_layout(tag_level = 'new')
pfail[[3]] <- pfail[[3]] + plot_layout(tag_level = 'new')

pfail <- pfail + plot_annotation(tag_levels = c('I', 'a'), tag_sep = '.', tag_suffix = ':') +
  plot_layout(guides = 'collect')

print(pfail)

曾尝试用plot_spacer()添加间距，但处理约20张图时问题更严重，寻求有效解决办法。

1. 抵消嵌套布局的间距叠加 #

问题根源是嵌套的plot_layout会默认叠加面板间距，导致分组间的间距翻倍。可以在分组布局中设置gap = unit(0, "cm")，只在顶层布局统一控制间距：

# 分组时关闭子布局的间距
pA <- p1 | p2 + plot_layout(axis_titles = 'collect', gap = unit(0, "cm"))
pB <- p3 | p4 + plot_layout(axis_titles = 'collect', gap = unit(0, "cm"))
pC <- p5 | p6 + plot_layout(axis_titles = 'collect', gap = unit(0, "cm"))

# 顶层布局统一设置全局间距
pfixed <- pA | pB | pC +
  plot_layout(guides = 'collect', gap = unit(0.5, "cm")) +
  plot_annotation(tag_levels = c('I', 'a'), tag_sep = '.', tag_suffix = ':')

2. 使用wrap_plots统一管理所有子图 #

放弃先分组再组合的方式，直接用wrap_plots将所有子图按规则排列，避免多层嵌套带来的布局混乱，尤其适合处理20张这类大量绘图的场景：

pfixed <- wrap_plots(p1, p2, p3, p4, p5, p6, nrow = 3, ncol = 2) +
  plot_layout(axis_titles = 'collect', guides = 'collect', gap = unit(0.5, "cm")) +
  plot_annotation(tag_levels = list(c('I', 'II', 'III'), c('a', 'b')), tag_sep = '.', tag_suffix = ':')

3. 全局统一面板间距 #

给所有子图设置一致的theme(panel.spacing)，从源头避免间距不一致：

# 给每个子图添加统一的面板间距
base_theme <- theme(panel.spacing = unit(0.5, "cm"))

p1 <- ggplot(mpg, aes(class, cyl, group = manufacturer)) +
  geom_bar(stat = "identity", position = position_dodge()) + base_theme
p2 <- ggplot(mpg, aes(class, cyl, group = manufacturer)) +
  geom_line(aes(linetype = manufacturer, color = manufacturer), position = position_dodge2(width = 0.2), linewidth = 0.5) + base_theme
# p3-p6同理添加base_theme

# 组合时顶层布局关闭额外间距
pfixed <- pA | pB | pC +
  plot_layout(guides = 'collect', gap = unit(0, "cm")) +
  plot_annotation(tag_levels = c('I', 'a'), tag_sep = '.', tag_suffix = ':')

内容的提问来源于stack exchange，提问作者Yeahhiam