I want to print list of graphs throw .Rnw. Wonder how to accomplish the task.
\documentclass{article}
\begin{document}
\SweaveOpts{concordance=TRUE}
\begin{figure}[H]
<< label = Plot >>=
library(tidyverse)
Ys <- c("disp", "hp", "drat", "wt", "qsec")
Plots <-
Ys %>%
map(function(y)
ggplot(data = mtcars, mapping = aes(x = mpg, y = y)) +
geom_point(aes_string(y = y))
)
print(Plots)
@
\end{figure}
\end{document}
You can use one of the variations on purrr::walk to silently map over the list of plots and save them with ggsave. walk2 maps over two lists, or iwalk maps over a list and its names. iwalk is really simple if you give the plot list names; the way I did it here is a little awkward, but in the wild, such as when plotting from a data frame that's been split, or with a vector that already has names, it could be different.
aes_string is used for creating aes calls programmatically, but it's being deprecated in favor of tidy eval; one example of this is in the aes docs, but I also was reminded of how to do it by this answer.
I first name the vector, with its own values as its names. Then I map over it and apply a function that takes a string y and creates a symbol of it using rlang::sym. This allows me to use tidy eval to create a bare column name with !!y_sum.
# https://stackoverflow.com/a/50726130/5325862
Plots <- Ys %>%
setNames(., nm = .) %>%
map(function(y) {
y_sym <- rlang::sym(y)
ggplot(data = mtcars, aes(x = mpg, y = !!y_sym)) +
geom_point()
})
Example of one plot in the list, which can be accessed by name:
Plots$hp
Then to save the plots, iwalk over using the names to create filenames. You can add additional parameters to ggsave, such as dimensions and resolution.
Plots %>%
iwalk(~ggsave(filename = sprintf("%s_vs_mpg.png", .y), plot = .x))
#> Saving 7 x 5 in image
#> Saving 7 x 5 in image
#> Saving 7 x 5 in image
#> Saving 7 x 5 in image
#> Saving 7 x 5 in image
You can do everything in one step if you need: map along to make plots and save in a single function. I went with the approach of saving the plots to a list, since that's what you did in the question, and that allows you to pull up individual plots in your document, rather than only saving them as files.
This will create the five pngs in the working directory, (which is the location of the Rnw file, unless you've modified it with something like opts_knit$set(root.dir='../../')).
Ys %>%
purrr::iwalk(
~ggplot(mtcars, aes_string(x = "mpg", y = .x)) +
geom_point() +
ggsave(filename = .y, device = "png")
)
try:
\documentclass{article}
\begin{document}
\SweaveOpts{concordance=TRUE}
<< label = Plot >>=
library(tidyverse)
Ys <- c("disp", "hp", "drat", "wt", "qsec")
Plots <-
Ys %>%
map(function(y)
ggplot(data = mtcars, mapping = aes(x = mpg, y = y)) +
geom_point(aes_string(y = y))
)
@
<<results=tex,echo=FALSE, >>=
for(i in 1:length(Plots)){
file=paste("myfile", i, ".png", sep="")
png(file, res = 80)
print(Plots[[i]])
dev.off()
cat( "\\begin{figure}\n\\includegraphics{", file, "}\n\n\\end{figure}\n", sep="")
}
@
\end{document}
knitr\documentclass{article}
\begin{document}
\SweaveOpts{concordance=TRUE}
\begin{figure}
<< label = Plot , fig=TRUE,results=hide>>=
library(tidyverse)
library(gridExtra)
Ys <- c("disp", "hp", "drat", "wt", "qsec")
plotLlist<-list()
for (i in 1:length(Ys)) {
plotLlist[[i]]<- ggplot(data = mtcars, aes_string(x = "mpg", y = Ys[i])) +
geom_point(aes_string(y = Ys[i]))
}
print(grid.arrange(grobs =plotLlist[1:4], nrow = 2))
@
\end{figure}
\end{document}
