如何在ggplot分面面板中保持条形宽度一致？

使用ggplot2绘制带facet_wrap的条形图时，因不同面板的x轴类别数量不同，导致条形宽度不一致，需要修改代码使所有面板的条形宽度保持统一。

原始数据及绘图代码如下：

set.seed(123) # For reproducibility
df_long <- data.frame(
 trait = rep(c("Trait1", "Trait2", "Trait3", "Trait4", "Trait5"), each = 4), 
 type = rep(c("eur_ita", "eur_ita_2", "eur_ita_3", "eur_ita_4"), times = 5),   
 value = c(runif(5, 5, 15), runif(5, 8, 20), runif(5, 1, 12), runif(5, 4, 18)),
 macro_cat3 = rep(c("Category1", "Category2", "Category3"), each = 10) # Dividing into two macro categories
)

ggplot(df_long, aes(x = trait, y = value, fill = type, alpha = type)) +
  geom_col(position = position_dodge(width = 0.6), color = "black", width = 0.6) +  # Adjust dodge and bar width
  facet_wrap(~ macro_cat3, scales = "free_x", ncol = 4) +  # Keep free_x scales
  labs(x = "Trait", y = "Percentage of r2", fill = " Type") +
  scale_fill_manual(
    values = c(
      "eur_ita" = "red",
      "eur_ita_2" = "black",
      "eur_ita_3" = "green",
      "eur_ita_4" = "gray"
    ),
    name = NULL  # Removes the legend title
  ) +
  scale_alpha_manual(
    values = c(
      "eur_ita" = 1,
      "eur_ita_2" = 0.9,
      "eur_ita_3" = 0.9,
      "eur_ita_4" = 1
    ),
    guide = "none"
  ) 

方法1：用ggh4x包固定面板宽度 #

ggh4x扩展包的facet_wrap2函数支持fixed_width参数，可强制所有面板使用相同宽度，从而保证条形宽度一致。

修改后的代码：

# 首次使用需安装包
# install.packages("ggh4x")
library(ggh4x)

ggplot(df_long, aes(x = trait, y = value, fill = type, alpha = type)) +
  geom_col(position = position_dodge(width = 0.6), color = "black", width = 0.6) +
  # 替换facet_wrap为facet_wrap2，设置固定宽度
  facet_wrap2(~ macro_cat3, scales = "free_x", ncol = 4, fixed_width = unit(2, "cm")) +
  labs(x = "Trait", y = "Percentage of r2", fill = " Type") +
  scale_fill_manual(
    values = c(
      "eur_ita" = "red",
      "eur_ita_2" = "black",
      "eur_ita_3" = "green",
      "eur_ita_4" = "gray"
    ),
    name = NULL
  ) +
  scale_alpha_manual(
    values = c(
      "eur_ita" = 1,
      "eur_ita_2" = 0.9,
      "eur_ita_3" = 0.9,
      "eur_ita_4" = 1
    ),
    guide = "none"
  )

方法2：统一x轴因子水平 #

将trait转换为包含所有可能水平的因子，并设置drop=FALSE，让每个面板的x轴保留所有trait的位置（无数据的trait留空），以此保证x轴长度一致，条形宽度统一。

修改后的代码：

# 统一trait的因子水平
df_long$trait <- factor(df_long$trait, levels = c("Trait1", "Trait2", "Trait3", "Trait4", "Trait5"))

ggplot(df_long, aes(x = trait, y = value, fill = type, alpha = type)) +
  geom_col(position = position_dodge(width = 0.6), color = "black", width = 0.6) +
  facet_wrap(~ macro_cat3, scales = "free_x", ncol = 4) +
  labs(x = "Trait", y = "Percentage of r2", fill = " Type") +
  scale_x_discrete(drop = FALSE) + # 保留所有因子水平，避免空trait被删除
  scale_fill_manual(
    values = c(
      "eur_ita" = "red",
      "eur_ita_2" = "black",
      "eur_ita_3" = "green",
      "eur_ita_4" = "gray"
    ),
    name = NULL
  ) +
  scale_alpha_manual(
    values = c(
      "eur_ita" = 1,
      "eur_ita_2" = 0.9,
      "eur_ita_3" = 0.9,
      "eur_ita_4" = 1
    ),
    guide = "none"
  )

注：方法2会在无对应trait数据的面板显示空白标签位，若需隐藏空白标签，可额外调整主题的x轴标签样式。

内容的提问来源于stack exchange，提问作者Khaleesi95