ggplot2基于tabyl绘制geom_col时图例与条形顺序匹配方法

代码出现排序、图例错乱的核心原因有两点：

1. 硬编码Category因子水平时没有和百分比计算结果绑定，后续fill映射嵌套str_wrap()会把因子型变量转成普通字符向量，之前手动设置的因子水平直接失效。ggplot对字符型分类变量默认按字母顺序排序，这就是图例始终按字母排列的核心原因。
2. arrange()仅调整数据框行顺序，不会修改因子本身的水平顺序。水平条形图的y轴默认从下到上按因子水平的先后顺序排列，没有同步更新因子水平的前提下，仅排序数据框会直接导致条形顺序和预期相反。

条形顺序稳定控制 #

不要手动硬编码长长的类别水平，完成百分比降序排列后，直接用forcats包的函数按数据排序结果设置因子水平即可：

• 调用fct_inorder()把因子水平设置为当前数据的行顺序
• 如果需要百分比最高的条形显示在y轴视觉最上方，再接一层fct_rev()反转水平顺序即可，一次设置永久生效，后续所有几何对象（条形、误差线、标签）的位置都会自动对齐。

图例顺序对齐 #

不要在aes()内部做变量转换（比如套str_wrap()），这类转换会丢失因子的排序属性。文本换行的操作放到scale_fill_*的labels参数中实现，只要fill映射的是排好序的因子变量，图例顺序会自动和条形顺序完全匹配，不需要手动硬编码limits。

library(tidyverse)
library(ggplot2)
library(forcats)
library(janitor)

temp <- tribble(
  ~ Category,
  "AAAAAAA AAAAAAAA AAAAAAAAA AAAAAAAAAA AAAAAAAAAAA AAAAAAAAAAA AAAAAAAA AAAAAAAAAA",
  "AAAAAAA AAAAAAAA AAAAAAAAA AAAAAAAAAA AAAAAAAAAAA AAAAAAAAAAA AAAAAAAA AAAAAAAAAA",
  "AAAAAAA AAAAAAAA AAAAAAAAA AAAAAAAAAA AAAAAAAAAAA AAAAAAAAAAA AAAAAAAA AAAAAAAAAA",
  "AAAAAAA AAAAAAAA AAAAAAAAA AAAAAAAAAA AAAAAAAAAAA AAAAAAAAAAA AAAAAAAA AAAAAAAAAA",
  "AAAAAAA AAAAAAAA AAAAAAAAA AAAAAAAAAA AAAAAAAAAAA AAAAAAAAAAA AAAAAAAA AAAAAAAAAA",
  "AAAAAAA AAAAAAAA AAAAAAAAA AAAAAAAAAA AAAAAAAAAAA AAAAAAAAAAA AAAAAAAA AAAAAAAAAA",
  "AAAAAAA AAAAAAAA AAAAAAAAA AAAAAAAAAA AAAAAAAAAAA AAAAAAAAAAA AAAAAAAA AAAAAAAAAA",
  "AAAAAAA AAAAAAAA AAAAAAAAA AAAAAAAAAA AAAAAAAAAAA AAAAAAAAAAA AAAAAAAA AAAAAAAAAA",
  "AAAAAAA AAAAAAAA AAAAAAAAA AAAAAAAAAA AAAAAAAAAAA AAAAAAAAAAA AAAAAAAA AAAAAAAAAA",
  "AAAAAAA AAAAAAAA AAAAAAAAA AAAAAAAAAA AAAAAAAAAAA AAAAAAAAAAA AAAAAAAA AAAAAAAAAA",
  "BBBB BBBB BBBBB BBBBBBBBB BBBBBB BBBBB BBBBBB BBBBB BBBBB BBBBBBBBB BBBB BBBBB BBBBBBB",
  "BBBB BBBB BBBBB BBBBBBBBB BBBBBB BBBBB BBBBBB BBBBB BBBBB BBBBBBBBB BBBB BBBBB BBBBBBB",
  "BBBB BBBB BBBBB BBBBBBBBB BBBBBB BBBBB BBBBBB BBBBB BBBBB BBBBBBBBB BBBB BBBBB BBBBBBB",
  "BBBB BBBB BBBBB BBBBBBBBB BBBBBB BBBBB BBBBBB BBBBB BBBBB BBBBBBBBB BBBB BBBBB BBBBBBB",
  "BBBB BBBB BBBBB BBBBBBBBB BBBBBB BBBBB BBBBBB BBBBB BBBBB BBBBBBBBB BBBB BBBBB BBBBBBB",
  "BBBB BBBB BBBBB BBBBBBBBB BBBBBB BBBBB BBBBBB BBBBB BBBBB BBBBBBBBB BBBB BBBBB BBBBBBB",
  "BBBB BBBB BBBBB BBBBBBBBB BBBBBB BBBBB BBBBBB BBBBB BBBBB BBBBBBBBB BBBB BBBBB BBBBBBB",
  "BBBB BBBB BBBBB BBBBBBBBB BBBBBB BBBBB BBBBBB BBBBB BBBBB BBBBBBBBB BBBB BBBBB BBBBBBB",
  "BBBB BBBB BBBBB BBBBBBBBB BBBBBB BBBBB BBBBBB BBBBB BBBBB BBBBBBBBB BBBB BBBBB BBBBBBB",
  "BBBB BBBB BBBBB BBBBBBBBB BBBBBB BBBBB BBBBBB BBBBB BBBBB BBBBBBBBB BBBB BBBBB BBBBBBB",
  "BBBB BBBB BBBBB BBBBBBBBB BBBBBB BBBBB BBBBBB BBBBB BBBBB BBBBBBBBB BBBB BBBBB BBBBBBB",
  "BBBB BBBB BBBBB BBBBBBBBB BBBBBB BBBBB BBBBBB BBBBB BBBBB BBBBBBBBB BBBB BBBBB BBBBBBB",
  "CCCCC CCC CCC CC CCCCC CCC CCCCCCCCCC CCCC CCCCC CCCCCCCCC CCCCCCCCCCC CCCC CCC CCC C CCC",
  "CCCCC CCC CCC CC CCCCC CCC CCCCCCCCCC CCCC CCCCC CCCCCCCCC CCCCCCCCCCC CCCC CCC CCC C CCC",
  "CCCCC CCC CCC CC CCCCC CCC CCCCCCCCCC CCCC CCCCC CCCCCCCCC CCCCCCCCCCC CCCC CCC CCC C CCC",
  "CCCCC CCC CCC CC CCCCC CCC CCCCCCCCCC CCCC CCCCC CCCCCCCCC CCCCCCCCCCC CCCC CCC CCC C CCC",
  "DDDDD DD D DDD DDDD DDD DDDDDDD DDD DDDD DDDDDDD DDD DDD DDDD DDDDDDDDD DDDD DDDDD DDDDDDD",
  "DDDDD DD D DDD DDDD DDD DDDDDDD DDD DDDD DDDDDDD DDD DDD DDDD DDDDDDDDD DDDD DDDDD DDDDDDD",
  "DDDDD DD D DDD DDDD DDD DDDDDDD DDD DDDD DDDDDDD DDD DDD DDDD DDDDDDDDD DDDD DDDDD DDDDDDD",
  "DDDDD DD D DDD DDDD DDD DDDDDDD DDD DDDD DDDDDDD DDD DDD DDDD DDDDDDDDD DDDD DDDDD DDDDDDD",
  "DDDDD DD D DDD DDDD DDD DDDDDDD DDD DDDD DDDDDDD DDD DDD DDDD DDDDDDDDD DDDD DDDDD DDDDDDD",
  "DDDDD DD D DDD DDDD DDD DDDDDDD DDD DDDD DDDDDDD DDD DDD DDDD DDDDDDDDD DDDD DDDDD DDDDDDD",
  "DDDDD DD D DDD DDDD DDD DDDDDDD DDD DDDD DDDDDDD DDD DDD DDDD DDDDDDDDD DDDD DDDDD DDDDDDD",
  "DDDDD DD D DDD DDDD DDD DDDDDDD DDD DDDD DDDDDDD DDD DDD DDDD DDDDDDDDD DDDD DDDDD DDDDDDD",
  "DDDDD DD D DDD DDDD DDD DDDDDDD DDD DDDD DDDDDDD DDD DDD DDDD DDDDDDDDD DDDD DDDDD DDDDDDD",
  "DDDDD DD D DDD DDDD DDD DDDDDDD DDD DDDD DDDDDDD DDD DDD DDDD DDDDDDDDD DDDD DDDDD DDDDDDD",
  "DDDDD DD D DDD DDDD DDD DDDDDDD DDD DDDD DDDDDDD DDD DDD DDDD DDDDDDDDD DDDD DDDDD DDDDDDD",
  "DDDDD DD D DDD DDDD DDD DDDDDDD DDD DDDD DDDDDDD DDD DDD DDDD DDDDDDDDD DDDD DDDDD DDDDDDD",
  "DDDDD DD D DDD DDDD DDD DDDDDDD DDD DDDD DDDDDDD DDD DDD DDDD DDDDDDDDD DDDD DDDDD DDDDDDD",
  "DDDDD DD D DDD DDDD DDD DDDDDDD DDD DDDD DDDDDDD DDD DDD DDDD DDDDDDDDD DDDD DDDDD DDDDDDD",
  "DDDDD DD D DDD DDDD DDD DDDDDDD DDD DDDD DDDDDDD DDD DDD DDDD DDDDDDDDD DDDD DDDDD DDDDDDD",
  "DDDDD DD D DDD DDDD DDD DDDDDDD DDD DDDD DDDDDDD DDD DDD DDDD DDDDDDDDD DDDD DDDDD DDDDDDD",
  "EEEE",
  "EEEE",
  "EEEE",
  "EEEE",
)

temp_n <- temp %>% nrow()

temp_tabyl <- 
  temp %>% 
  tabyl(Category) %>% 
  rename(Percent = percent) %>% 
  arrange(desc(Percent)) %>% 
  # 核心修改：按排序结果设置因子，反转顺序让最大值在y轴最上方
  mutate(Category = fct_inorder(Category) %>% fct_rev(),
         CI = sqrt(Percent*(1-Percent)/temp_n),
         MOE = CI * 1.96,
         ub = Percent + MOE,
         lb = Percent - MOE)

temp_tabyl %>% 
  ggplot() + 
  geom_col(aes(y = Category, 
               x = Percent,
               fill = Category), # 直接映射排好序的因子，不要在这里套str_wrap
           colour = "black"
  ) + 
  geom_errorbar(
    aes(
      y = Category,
      xmin = lb,
      xmax = ub
    ),
    width = 0.4,
    colour = "orange",
    alpha = 0.9,
    size = 1.3
  ) + 
  geom_label(aes(y = Category, 
                 x = Percent,
                 label = scales::percent(Percent)),nudge_x = .11) + 
  scale_x_continuous(labels = scales::percent,limits = c(0,1)) + 
  # 在这里统一处理图例文本换行，不丢失因子排序属性
  scale_fill_discrete(labels = ~str_wrap(.x, 40), name = "Category") +
  labs(title = "Plot Title",
       caption = "Plot Caption.") +
  theme_bw() +
  theme(
    text = element_text(family = 'Roboto'),
    strip.text.x = element_text(size = 14,
                                face = 'bold'),
    panel.grid.minor = element_blank(),
    axis.title.y = element_text(size = 14),
    plot.title = element_text(hjust = 0.5, size = 16),
    plot.subtitle = element_text(hjust = 1),
    plot.caption = element_text(hjust = 0),
    axis.text.y=element_blank()
  ) +
  theme(panel.grid.major = element_blank(),
        panel.grid.minor = element_blank()) +
  theme(strip.text = element_text(colour = 'white'),
        legend.spacing.y = unit(.5, 'cm')) + 
  guides(fill = guide_legend(byrow = TRUE))

运行后即可实现：条形按百分比从大到小从上到下排列、每个类别独立填充色、隐藏y轴刻度标签、图例顺序和条形顺序完全一致，误差线、数值标签位置全部自动对齐。

内容的提问来源于stack exchange，提问作者Abe