如何合并两个ggplot2自定义函数并使图例独立区分？

问题根源在于两个自定义函数各自重复添加了scale_alpha_manual，后调用的函数会覆盖前者的刻度设置；同时你手动指定的color和linetype未关联到图例映射，导致图例无法正确显示两条线的样式差异。以下是两种可行的解决方案：

方案1：保留alpha映射，统一设置图例样式 #

先修改自定义函数，移除每个函数内的labs和scale_alpha_manual（避免重复覆盖）：

plt_regression_line <- function(color = "black", linetype = "dashed"){
  list(
    ggplot2::geom_smooth(
      method = "lm",
      se = FALSE,
      formula = "y ~ x",
      mapping = ggplot2::aes(alpha = "Regression line"),
      linetype = linetype,
      color = color)
  )
}

plt_identity_line <- function(color = "black", linetype = "dashed"){
  list(
    ggplot2::geom_abline(
      mapping = ggplot2::aes(slope = 1, intercept = 0, alpha = "Identity line"),
      linetype = linetype,
      color = color)
  )
}

然后在主绘图代码中统一设置图例，通过guide_legend的override.aes参数指定每个图例条目的颜色和线型：

library(ggplot2)

ggplot(mtcars,aes(qsec,mpg))+
  geom_point()+
  plt_regression_line(color = "red",linetype = 1)+
  plt_identity_line(color = "blue",linetype = 1)+
  labs(alpha = "")+  # 清空图例标题
  scale_alpha_manual(values = c(1,1))+  # 保持线条不透明
  guides(alpha = guide_legend(override.aes = list(
    color = c("red", "blue"),  # 对应两条线的颜色
    linetype = c(1,1)          # 对应两条线的线型
  )))

方案2：改用linetype映射（更直观，不占用alpha） #

如果不需要用alpha做其他映射，推荐改用linetype区分线条，更符合ggplot2的映射逻辑：
修改自定义函数，将映射从alpha改为linetype：

plt_regression_line <- function(color = "black", linetype = "dashed"){
  list(
    ggplot2::geom_smooth(
      method = "lm",
      se = FALSE,
      formula = "y ~ x",
      mapping = ggplot2::aes(linetype = "Regression line"),
      color = color)
  )
}

plt_identity_line <- function(color = "black", linetype = "dashed"){
  list(
    ggplot2::geom_abline(
      mapping = ggplot2::aes(slope = 1, intercept = 0, linetype = "Identity line"),
      color = color)
  )
}

主绘图代码中统一设置图例：

ggplot(mtcars,aes(qsec,mpg))+
  geom_point()+
  plt_regression_line(color = "red",linetype = 1)+
  plt_identity_line(color = "blue",linetype = 1)+
  labs(linetype = "")+
  scale_linetype_manual(values = c(1,1))+  # 统一线型为实线
  guides(linetype = guide_legend(override.aes = list(color = c("red", "blue"))))

内容的提问来源于stack exchange，提问作者Vinícius Félix