如何合并R基础绘图、lattice与ggplot2三种系统生成的图形？

嘿，这个问题我之前也碰到过——毕竟R里三个绘图系统各有拿手好戏，但要把它们凑到同一张图里确实得费点心思。下面给你几个实用的方案，亲测好用：

首先先把你给出的三个绘图代码优化一下（尽量不用attach()，避免环境变量混乱）：

• 基础绘图箱线图：

boxplot(mpg~cyl, data = mtcars, 
        xlab = "Number of Cylinders", ylab = "Miles per Gallon")

• Lattice的bwplot：

library(lattice)
bwplot(~mpg | cyl, data = mtcars, 
       xlab = "Number of Cylinders", ylab = "Miles per Gallon")

• ggplot2的箱线图（补全你的qplot代码）：

library(ggplot2)
mtcars$cyl <- as.factor(mtcars$cyl)
qplot(cyl, mpg, data = mtcars, geom = "boxplot", 
      xlab = "Number of Cylinders", ylab = "Miles per Gallon")

基础绘图用的是base graphics系统，而Lattice和ggplot2基于grid graphics，两者渲染逻辑不同，gridBase包能帮我们在两种系统间切换视口，实现混排。

示例代码（把三个图排成一行）：

# 加载依赖包
library(grid)
library(gridBase)
library(lattice)
library(ggplot2)

# 预处理数据
mtcars$cyl <- as.factor(mtcars$cyl)

# 设置整体布局：1行3列
layout(matrix(1:3, nrow = 1))

# 第一个图：基础绘图
par(mar = c(4, 4, 1, 1)) # 调整边距，避免重叠
boxplot(mpg~cyl, data = mtcars, 
        xlab = "Number of Cylinders", ylab = "Miles per Gallon")

# 第二个图：Lattice绘图
# 获取基础绘图的视口，切换到grid环境
vps <- baseViewports()
pushViewport(vps$inner, vps$figure, vps$plot)
bwplot(~mpg | cyl, data = mtcars, 
       xlab = "Number of Cylinders", ylab = "Miles per Gallon")
upViewport(3) # 退出视口

# 第三个图：ggplot2绘图
pushViewport(vps$inner, vps$figure, vps$plot)
# 关键：用print()并设置newpage=FALSE，避免ggplot新开一页
print(qplot(cyl, mpg, data = mtcars, geom = "boxplot", 
            xlab = "Number of Cylinders", ylab = "Miles per Gallon"),
      newpage = FALSE)
upViewport(3)

cowplot包专门解决多图排列问题，支持把base/Lattice/ggplot2的图统一转换成grid对象，然后一键排列，代码最简洁。

示例代码：

library(cowplot)
library(lattice)
library(ggplot2)

# 预处理数据
mtcars$cyl <- as.factor(mtcars$cyl)

# 1. 捕获基础绘图的内容
par(mar = c(4, 4, 1, 1))
boxplot(mpg~cyl, data = mtcars, 
        xlab = "Number of Cylinders", ylab = "Miles per Gallon")
base_plot <- recordPlot() # 把当前base图记录下来

# 2. 生成Lattice和ggplot2的图对象
lattice_plot <- bwplot(~mpg | cyl, data = mtcars, 
                       xlab = "Number of Cylinders", ylab = "Miles per Gallon")
ggplot_plot <- qplot(cyl, mpg, data = mtcars, geom = "boxplot", 
                     xlab = "Number of Cylinders", ylab = "Miles per Gallon")

# 3. 转换为grid兼容对象并排列
base_grob <- as_grob(base_plot) # 把base图转成grob
lattice_grob <- as_grob(lattice_plot) # Lattice图转grob

# 排列三个图，nrow设置行数，labels加标签区分
plot_grid(base_grob, lattice_grob, ggplot_plot, 
          nrow = 1, labels = c("Base Plot", "Lattice Plot", "ggplot2 Plot"))

gridExtra也是常用的多图排列工具，和cowplot思路类似，需要手动转换base图为grid对象：

library(gridExtra)
library(grid)
library(lattice)
library(ggplot2)

# 预处理数据
mtcars$cyl <- as.factor(mtcars$cyl)

# 1. 捕获并转换基础绘图
par(mar = c(4, 4, 1, 1))
boxplot(mpg~cyl, data = mtcars, 
        xlab = "Number of Cylinders", ylab = "Miles per Gallon")
base_grob <- grid.grabExpr(recordPlot()) # 捕获base图为grob

# 2. 转换Lattice和ggplot2图
lattice_grob <- as_grob(bwplot(~mpg | cyl, data = mtcars, 
                               xlab = "Number of Cylinders", ylab = "Miles per Gallon"))
ggplot_grob <- ggplotGrob(qplot(cyl, mpg, data = mtcars, geom = "boxplot", 
                                xlab = "Number of Cylinders", ylab = "Miles per Gallon"))

# 3. 排列图形
grid.arrange(base_grob, lattice_grob, ggplot_grob, nrow = 1)

小提示：尽量避免使用attach()，直接在绘图函数里用data=参数指定数据集，能减少很多环境变量的混乱问题。

内容的提问来源于stack exchange，提问作者nalzok