如何让ggplot分面图的Y轴标签沿公共中线对齐？

当使用facet_grid(scales = "free_y")时，element_text(hjust = .5)只能让每个分面内的Y轴标签沿各自分面的中线对齐，无法实现所有标签沿全局公共中线对齐。以下是两种可行的解决方法：

方法一：通过gtable手动调整布局 #

直接操作ggplot生成的gtable对象，统一所有Y轴标签的宽度，让它们基于全局最大宽度居中：

library(tidyverse)
library(gtable)
library(grid)

# 生成示例数据
t1 <- tibble(
  labs = c(
    "A short label",
    "Second short label",
    "This is a medium length label",
    "This is also a medium length",
    "Now, this label is actually quite long, really long",
    "This label is also actually quite long, really long"
  ) %>% 
    fct_inorder %>% 
    fct_rev,
  facs = str_c("facet", 1:3) %>% rep(each = 2),
  x = runif(6)
)

# 绘制基础图形
p <- t1 %>% 
  ggplot(aes(x, labs)) +
  geom_point() +
  facet_grid(vars(facs), scales = "free_y") +
  labs(x = "", y = "") +
  theme(axis.text.y = element_text(hjust = .5))

# 提取图形的gtable结构
gt <- ggplotGrob(p)

# 定位所有左侧Y轴文本的grobs
y_axis_grobs <- gt$grobs[grepl("axis-l", gt$layout$name)]

# 计算所有Y轴标签的最大宽度
max_label_width <- max(sapply(y_axis_grobs, function(g) g$widths[2]))

# 统一每个分面的Y轴标签宽度
for (i in seq_along(y_axis_grobs)) {
  y_axis_grobs[[i]]$widths[2] <- max_label_width
  gt$grobs[grepl("axis-l", gt$layout$name)][[i]] <- y_axis_grobs[[i]]
}

# 绘制调整后的图形
grid.draw(gt)

方法二：用patchwork拼接独立分面 #

将每个分面拆分为单独的ggplot对象，再用patchwork拼接并对齐Y轴，天然实现全局标签居中：

library(tidyverse)
library(patchwork)

# 生成示例数据
t1 <- tibble(
  labs = c(
    "A short label",
    "Second short label",
    "This is a medium length label",
    "This is also a medium length",
    "Now, this label is actually quite long, really long",
    "This label is also actually quite long, really long"
  ) %>% 
    fct_inorder %>% 
    fct_rev,
  facs = str_c("facet", 1:3) %>% rep(each = 2),
  x = runif(6)
)

# 为每个分面单独创建图形
facet_plots <- map(unique(t1$facs), function(facet_name) {
  t1 %>% 
    filter(facs == facet_name) %>% 
    ggplot(aes(x, labs)) +
    geom_point() +
    labs(x = "", y = "") +
    ggtitle(facet_name) +
    theme(axis.text.y = element_text(hjust = .5))
})

# 垂直拼接图形并对齐Y轴
wrap_plots(facet_plots, ncol = 1, align = "h")

方法对比 #

• gtable方法：基于原生ggplot对象调整，无需拆分图形，适合需要保留原始分面结构的场景，但需要对gtable布局有一定了解。
• patchwork方法：代码更简洁直观，拼接逻辑清晰，适合分面数量较少的场景，对齐效果更稳定。

内容的提问来源于stack exchange，提问作者tomw