R ggplot柱状图仅高亮Prime分类时无该值全高亮问题求助

当绘制的目标列不存在「Prime」值时，ifelse判断返回的填充类别只有Normal一种，ggplot默认会将调色板的第一个颜色映射给仅有的类别，刚好对应预设的高亮色，就出现了所有柱体全高亮的错误。

显式指定填充色的固定映射关系，不管Highlighted类别是否存在，都绑定固定的颜色值，避免默认调色板的自动匹配错误。只需在每个ggplot语句中添加scale_fill_manual()配置即可。

library(ggplot2)
library(gridExtra)

test.df <- data.frame(
  replace_punct = c("Prime", "Not Measured", "Not Measured", "Not Measured", "Not Measured", "Not Measured", "Not Measured", "Not Measured"), 
  pos = c("Measured", "Not Measured", "Not Measured", "Not Measured", "Not Measured", "Not Measured", "Not Measured", "Not Measured"), 
  spelling_corr = c("Measured", "Not Measured", "Not Measured", "Not Measured", "Not Measured", "Not Measured", "Prime", "Prime"), 
  stringsAsFactors = FALSE
)

# 公共填充色配置，可根据需求自行修改颜色值
fill_map <- c("Highlighted" = "#E41A1C", "Normal" = "#999999")

replace_punct_bar <- ggplot(data=test.df, aes(x = replace_punct, y = after_stat(count/sum(count)), fill = ifelse(replace_punct == "Prime", "Highlighted", "Normal"))) + 
  scale_y_continuous(labels = scales::percent) + 
  geom_bar() + 
  scale_fill_manual(values = fill_map) + # 新增固定填充映射
  labs(title = "Replace punctuation freq", y = "Percent", x = "Status") + 
  geom_hline(aes(yintercept=0.6), colour="#990000", linetype="dashed") + 
  theme(legend.position="none")

pos_bar <- ggplot(data=test.df, aes(x = pos, y = after_stat(count/sum(count)), fill = ifelse(pos == "Prime", "Highlighted", "Normal"))) + 
  scale_y_continuous(labels = scales::percent) + 
  geom_bar() + 
  scale_fill_manual(values = fill_map) + # 新增固定填充映射
  labs(title = "Part of speech freq", y = "Percent", x = "Status") + 
  geom_hline(aes(yintercept=0.6), colour="#990000", linetype="dashed") + 
  theme(legend.position="none")

spelling_corr_bar <- ggplot(data=test.df, aes(x = spelling_corr, y = after_stat(count/sum(count)), fill = ifelse(spelling_corr == "Prime", "Highlighted", "Normal"))) + 
  scale_y_continuous(labels = scales::percent) + 
  geom_bar() + 
  scale_fill_manual(values = fill_map) + # 新增固定填充映射
  labs(title = "Spelling correction freq", y = "Percent", x = "Status") + 
  geom_hline(aes(yintercept=0.6), colour="#990000", linetype="dashed") + 
  theme(legend.position="none")

grid.arrange(replace_punct_bar, pos_bar, spelling_corr_bar, ncol=3)

如果需要批量生成多个同类型的柱状图，可以封装为函数减少重复代码：

plot_bar <- function(col_name, title) {
  ggplot(data=test.df, aes(x = .data[[col_name]], y = after_stat(count/sum(count)), fill = ifelse(.data[[col_name]] == "Prime", "Highlighted", "Normal"))) + 
    scale_y_continuous(labels = scales::percent) + 
    geom_bar() + 
    scale_fill_manual(values = fill_map) +
    labs(title = title, y = "Percent", x = "Status") + 
    geom_hline(aes(yintercept=0.6), colour="#990000", linetype="dashed") + 
    theme(legend.position="none")
}

# 调用函数生成三个图
replace_punct_bar <- plot_bar("replace_punct", "Replace punctuation freq")
pos_bar <- plot_bar("pos", "Part of speech freq")
spelling_corr_bar <- plot_bar("spelling_corr", "Spelling correction freq")

内容的提问来源于stack exchange，提问作者Haza