I have 12 plots generated by a loop and I want to plot them with 3 rows and 2 columns on one page (2 pages in total). I know how to do it in R
pdf("1.pdf") par(mfrow = c(3, 2)) for (i in 1:12) { x <- 1:10 y <- 2*x + rnorm(x) plot(x, y) } dev.off() But how to do this using ggplot?
library(ggplot2) library(grid) library(gridExtra) for (i in 1:12) { x <- 1:10 y <- 2*x + rnorm(x) qplot(x, y) } I think I have to use grid.arrange but need some help on that. Thanks.
You might want to take a look at the cowplot package that allows more flexibility than just using a naked grid.arrange.
This works - albeit a bit inelegantly:
library(ggplot2) library(grid) library(gridExtra) lg <- list() for (i in 1:12) { x <- 1:10 y <- 2*x + rnorm(x) lg[[i]] <- qplot(x, y) } grid.arrange(lg[[1]],lg[[2]],lg[[3]],lg[[4]],lg[[5]],lg[[6]],nrow=3,ncol=2) grid.arrange(lg[[7]],lg[[8]],lg[[9]],lg[[10]],lg[[11]],lg[[12]],nrow=3,ncol=2) Another more elegant but somewhat obtuse way to do the grid.arrange is the following (thanks to Axeman and beetroot - note the comments).
do.call(grid.arrange, c(lg[1:6], nrow = 3)) do.call(grid.arrange, c(lg[7:12], nrow = 3)) or this:
grid.arrange(grobs = lg[1:6], ncol=2) grid.arrange(grobs = lg[7:12], ncol=2) They all result in this - (think two of these - they look the same anyway):
do.call(grid.arrange, c(lg[1:6], nrow = 3)) and do.call(grid.arrange, c(lg[7:12], nrow = 3))
cowplot better.marrangeGrob is a convenient wrapper for multiple pages,
marrangeGrob(lg, ncol=2, nrow=3) or you can call grid.arrange() explicitly twice,
grid.arrange(grobs = lg[1:6], ncol=2) grid.arrange(grobs = lg[7:12], ncol=2) 