如何在ggplot中用两个数据集将折线图叠加到分组条形图上？

错误原因 #

你在全局图层的aes()中定义了fill = ReportAge，添加geom_line()时，它会默认继承这个全局美学映射，但你的dt2数据框中不存在ReportAge列，因此触发报错。

解决方法 #

有两种可行方案：

• 给geom_line()设置inherit.aes = FALSE，完全不继承全局美学映射
• 在geom_line()的aes()中显式覆盖fill映射（如fill = NULL）

同时注意：原代码中mutate(as.factor(ReportAge))未将转换结果赋值给列，会导致ReportAge仍为数值型，需修正为ReportAge = as.factor(ReportAge)。

修正后的完整代码 #

library(tidyverse)
library(lubridate)

dt1 <- expand.grid("ReportDate" = seq(ymd("2024-04-25"),
                                      ymd("2024-05-01"),
                                      by = "day"),
                   "ReportAge" = seq(1,7,1)) %>%
  mutate(n = sample(0:10, n(), replace = TRUE)) %>%
  arrange(ReportDate) %>%
  mutate(ReportAge = as.factor(ReportAge)) # 修正：赋值转换后的因子类型

dt2 <- dt1 %>%
  filter(as.numeric(ReportAge) > 5) %>%
  group_by(ReportDate) %>%
  summarize(OldReports = sum(n))

graph <- dt1 %>% 
  ggplot(aes(x = ReportDate, y = n, fill = ReportAge)) +
  geom_bar(stat = "identity",
           color = "black", 
           position = position_dodge2(width = 0.9, 
                                      preserve = "single")) +
  geom_text(aes(y= -1, label = ReportAge), 
            color = "black", 
            position = position_dodge2(width = 0.9, preserve = "single"),
            size = 3.5) +
  geom_text(aes(label = n),
            color = "red", 
            vjust = -0.3, 
            position = position_dodge2(width = 0.9, preserve = "single")) +
  guides(fill = guide_legend(override.aes = list(size = 0.5))) +
  scale_x_date(date_breaks = "day", date_labels = "%b %d") +
  theme_minimal()

# 方案1：禁用全局美学继承
graph2 <- graph +
  geom_line(data = dt2, mapping = aes(x = ReportDate, y = OldReports), 
            color = "red", inherit.aes = FALSE)

# 方案2：显式覆盖fill映射
# graph2 <- graph +
#   geom_line(data = dt2, mapping = aes(x = ReportDate, y = OldReports, fill = NULL), 
#             color = "red")

print(graph2)

内容的提问来源于stack exchange，提问作者Andrew Chastain