ggplot2绘图报错：Error in FUN(X[[i]], ...) : object 'a' not found 求助

你遇到的object 'a' not found报错，核心原因是ggplot的全局美学映射继承机制：
你最初创建绘图对象时，把y=a放在了全局的ggplot()函数里：

tpr <- ggplot(tpreg1, aes(x = t, y = a)) + ...

这意味着后续所有图层都会默认继承y=a的映射。但你添加geom_ribbon时用的是tpreg3数据，而tpreg3只有t列，没有a列，ggplot尝试调用这个不存在的列时自然就报错了。

方案1：把y=a映射移到需要它的图层（推荐） #

把y=a从全局ggplot()中移除，只在geom_point和geom_smooth这两个需要用到y=a的图层里单独指定，这样后续添加其他图层时就不会被不必要的映射干扰：

# 重新构建基础绘图，只保留所有图层共用的x=t映射
tpr <- ggplot(tpreg1, aes(x = t)) + 
  geom_point(aes(y = a), shape = 1, fill = "darkgreen", alpha = 1, color = "darkgreen", size = 1) + 
  geom_smooth(aes(y = a), method = lm, size = 0.5, linetype = "solid", color = "darkgreen")

# 后续的刻度范围、宽高比、主题设置保持不变
tpr <- tpr + expand_limits(x = c(21.5, 32.5), y = c(-16, 34)) + 
  scale_x_continuous(expand = c(0, 0), breaks = c(22,24,26,28,30,32)) + 
  scale_y_continuous(expand = c(0,0), breaks = c(-16,-8,0,8,16,24,32))
tpr<- tpr + theme(aspect.ratio = 0.75)
tpr <- tpr + theme(panel.background = element_blank()) + 
  theme(panel.grid.minor = element_blank(), panel.grid.major = element_blank())
tpr <- tpr + theme(axis.line.x = element_line(colour = "black", size = 0.5), 
                   axis.line.y = element_line(colour = "black", size = 0.5), 
                   axis.ticks.y = element_line(colour = "black", size = 0.5), 
                   axis.ticks.x = element_line(colour = "black", size = 0.5), 
                   axis.ticks.length = unit(.1, "cm"))

# 添加灰色面板，此时不会有映射冲突
tpr <- tpr + 
  geom_ribbon(data = tpreg3, aes(ymin = func1, ymax = func2), fill = "grey", alpha = .25) + 
  geom_ribbon(data = tpreg3, aes(ymin = func3, ymax = func4), fill = "grey", alpha = .25)

tpr # 现在可以正常生成带灰色面板的图

方案2：在geom_ribbon中禁用美学继承 #

如果你不想修改前面的代码，可以在geom_ribbon里加上inherit.aes = FALSE，明确告诉它不要继承全局的y=a映射：

# 直接修改添加灰色面板的代码
tpr <- tpr + 
  geom_ribbon(data = tpreg3, aes(x = t, ymin = func1, ymax = func2), fill = "grey", alpha = .25, inherit.aes = FALSE) + 
  geom_ribbon(data = tpreg3, aes(x = t, ymin = func3, ymax = func4), fill = "grey", alpha = .25, inherit.aes = FALSE)

两种方案都能解决报错，方案1更符合ggplot的最佳实践——只把所有图层共用的映射放在全局，专属映射放在对应图层里，能避免后续添加新图层时出现类似的继承冲突。

内容的提问来源于stack exchange，提问作者Kou