当我使用 ggnewscale 包创建多个比例时如何更改图例标题位置?

How to change legend title position when I have multiple scales created with ggnewscale package?

我创建了一个图,其中包含使用 ggnewscale 包制作的 2 个色标和 2 个线型标尺。我希望图例出现在图的底部,图例标题出现在标签上方。我做了一半,因为在我的尝试中,第一个音阶(来自 geom_vline())没有按照我想要的方式出现。在用 new_scale()new_scale_color() 声明下一个比例之前,我什至试图指出这个比例的所需位置。有人知道我该如何解决这个问题吗?

library(ggplot2)
library(ggnewscale)

mtcars$cyl <- factor(mtcars$cyl)

# Attempt 1

ggplot(data = mtcars,
       aes(x = wt,
           y = mpg)) +
  geom_vline(aes(xintercept = gear,
                 color = "Important line",
                 linetype = "Important line")) +
  scale_linetype_manual(name = "Important lines",
                        values = 1) +
  scale_color_manual(name = "Important lines",
                     values = "red") +
  new_scale("linetype") +
  new_scale_color() +
  geom_line(aes(linetype = cyl,
                color = cyl)) +
  theme(legend.position = "bottom") +
  guides(linetype = guide_legend(title.position = "top"),
         color = guide_legend(title.position = "top"))

# Attempt 2

ggplot(data = mtcars,
       aes(x = wt,
           y = mpg)) +
  geom_vline(aes(xintercept = gear,
                 color = "Important line",
                 linetype = "Important line")) +
  scale_linetype_manual(name = "Important lines",
                        values = 1) +
  scale_color_manual(name = "Important lines",
                     values = "red") +
  guides(linetype = guide_legend(title.position = "top",
                                 label.position = "bottom"), # Before new scales
         color = guide_legend(title.position = "top",
                              label.position = "bottom")) +
  new_scale("linetype") +
  new_scale_color() +
  geom_line(aes(linetype = cyl,
                color = cyl)) +
  theme(legend.position = "bottom") +
  guides(linetype = guide_legend(title.position = "top"),
         color = guide_legend(title.position = "top"))

两次尝试都产生相同的情节,这不是我想要的。

不确定是否还有其他选择,但在使用 ggnewscale 时对我有用的是通过比例尺的 guide 参数设置指南的样式。为了完成这项工作,我添加了一个´scale_color_discrete`:

library(ggplot2)
library(ggnewscale)

mtcars$cyl <- factor(mtcars$cyl)

ggplot(data = mtcars,
       aes(x = wt,
           y = mpg)) +
  geom_vline(aes(xintercept = gear,
                 color = "Important line",
                 linetype = "Important line")) +
  scale_linetype_manual(name = "Important lines",
                        values = 1, guide = guide_legend(title.position = "top") ) +
  scale_color_manual(name = "Important lines",
                     values = "red", guide = guide_legend(title.position = "top")) +
  new_scale("linetype") +
  new_scale_color() +
  geom_line(aes(linetype = cyl,
                color = cyl)) +
  scale_color_discrete(guide = guide_legend(title.position = "top")) +
  theme(legend.position = "bottom")