如何在ggplot2的facet_wrap子图中自定义X轴刻度标签？

问题 #

使用ggplot2的facet_wrap函数绘制分组堆叠柱状图时，通过scale_x_discrete设置X轴刻度标签后，所有标签在每个子图中重复显示。需求是为较长的刻度标签添加换行，同时避免标签在子图中重复。原代码如下：

phy_gre_5S_c0_samp = mean_data_type %>%
  filter(country == "Greece") %>% 
  filter(WP == "5S") %>% 
  filter(phase == "C0") %>% 
#  filter(Taxonomy != "Unclassified") %>% 
  ggplot(aes(x = sample_pretty_type,
             y = mean_relab,
             fill = Taxonomy)) + 
  geom_bar(stat = "identity",
           position = "stack") +
  theme(axis.text.x=element_blank(),
        axis.ticks.x=element_blank(),
        axis.title.x=element_blank(),
        legend.position="none") +
  scale_x_discrete(labels = c(c("Fertilized eggs","Yolk sack
larvae"), c("Inlet water
to tank/farm","Tank water","Waste water"), c("Distal intestine
tissue","Mid intestine tissue","Pyloric anterior
intestine tissue","Intestinal tissue"),c("Feed/diet"),c("Mid intestine
digesta","Posterior intestine
digesta", "Pyloric anterior
intestine digesta","Digesta"),c("Gills","Skin","Filet"))) +
  theme_classic() +
  labs(y = "Relative Abundance (%)",
       x = "Sample type") +
  scale_fill_manual(values = my_colors) +
  theme(axis.text.x = element_text(angle=45, vjust=1, hjust=1),
        legend.key.size = unit(2, 'cm'),
        legend.position = "right") +
  facet_wrap(vars(sample_category), scales = "free") +
  theme(text = element_text(size = 40))

解决方案 #

问题根源 #

直接用scale_x_discrete硬编码所有标签列表，会强制每个子图显示全部标签，忽略了facet_wrap的scales = "free"逻辑。正确做法是在数据预处理阶段就生成带换行的标签，让ggplot自动匹配子图对应的标签。

具体实现 #

1. 预处理数据，生成带换行的标签
   可以手动指定需要换行的标签，或用工具函数自动换行，避免硬编码全部标签。
2. 移除scale_x_discrete的手动标签设置
   让ggplot根据每个子图的实际数据自动加载对应标签。
3. 保留scales = "free"确保子图只显示自身的X轴标签

修改后的代码

# 预处理数据，添加带换行的标签列
mean_data_type_processed <- mean_data_type %>%
  filter(country == "Greece", WP == "5S", phase == "C0") %>%
  # 手动指定需要换行的标签，用\n表示换行
  mutate(sample_pretty_type = case_when(
    sample_pretty_type == "Yolk sack larvae" ~ "Yolk sack\nlarvae",
    sample_pretty_type == "Inlet water to tank/farm" ~ "Inlet water\nto tank/farm",
    sample_pretty_type == "Distal intestine tissue" ~ "Distal intestine\ntissue",
    sample_pretty_type == "Pyloric anterior intestine tissue" ~ "Pyloric anterior\nintestine tissue",
    sample_pretty_type == "Mid intestine digesta" ~ "Mid intestine\ndigesta",
    sample_pretty_type == "Posterior intestine digesta" ~ "Posterior intestine\ndigesta",
    sample_pretty_type == "Pyloric anterior intestine digesta" ~ "Pyloric anterior\nintestine digesta",
    # 其他标签保持原样
    TRUE ~ sample_pretty_type
  ))

# 绘制图表
phy_gre_5S_c0_samp <- mean_data_type_processed %>%
  ggplot(aes(x = sample_pretty_type,
             y = mean_relab,
             fill = Taxonomy)) + 
  geom_bar(stat = "identity", position = "stack") +
  theme_classic() +
  labs(y = "Relative Abundance (%)", x = "Sample type") +
  scale_fill_manual(values = my_colors) +
  theme(axis.text.x = element_text(angle=45, vjust=1, hjust=1),
        legend.key.size = unit(2, 'cm'),
        legend.position = "right",
        text = element_text(size = 40)) +
  facet_wrap(vars(sample_category), scales = "free")

更高效的自动换行方式

如果标签数量多，不想逐个指定，可以用stringr::str_wrap自动按宽度换行：

mean_data_type_processed <- mean_data_type %>%
  filter(country == "Greece", WP == "5S", phase == "C0") %>%
  mutate(sample_pretty_type = stringr::str_wrap(sample_pretty_type, width = 12))

内容的提问来源于stack exchange，提问作者djjohn8