R
From UC_Chemistry
Contents |
[edit]
Install R
- Download file from here and move it into your home directory. Then
./configure make
if you have root priviledge
make install
otherwise you don't have to install R; run it from the R-XXX/bin/ directory or create symbolic link in any place
ln -s /home/user-name/R-XXX/bin/R R
the link will be place in the current folder.
Without root priviledge you also have to export R_LIBS variable
export R_LIBS=/home/user-name/R-2.5.0/library
i think you have to write this line in ".bash_profile" file in your home directory
[edit]
Extra Packages
- To install a package, type in any place (provided you already can run R)
R CMD INSTALL package-name.tar.gz
- Most of the packages are at CRAN
- Bio3d is a must-have package for computational chemist
- Ade4 labels on 2D scatter plot
- R commander a GUI for windows (i think)
- XML xml parser
[edit]
Links
[edit]
R basic
- Load and save R objects
a <- c(1,2,3)
save(a, file="test.RData")
rm(a)
load("test.RData")
- Load data and plot XY
myplot <- read.table("file_with_two_columns.xy",header=T,sep=" ")
plot(myplot)
- Save plot to in ps-format
postscript("myfile.ps")
plot(myplot)
dev.off()
if you want to convert to epsi format, in command line type: system("ps2epsi myfile.ps")
- This example is marvelous, it will create 5 random normal distributions and will save in files fig1.jpg, fig2.jpg ..., fig5.jpg
for(i in 1:5) {
jpeg(paste("fig", i, ".jpg", sep = ""))
hist(rnorm(100))
dev.off()
}
Then in command line create animation (look Animated_Gifs for details)
convert -delay 50 -loop 50 fig*.jpg animated.gif
- Create 2 x 5 matrix with numbers from 1 to 10
x <- matrix(1:10,ncol=5)
[edit]
Scripts
- All R commands can be saved in a script, say foo.R and then executed as follows
R CMD BATCH foo.R &
[edit]
R packages
[edit]
Using R packages
Currently loaded packages
search()
See installed packages and load a package
library() library(packagename)
HTML documentation of all installed packages
help.start()
[edit]
Making R packages
- For instructions on making R packages see tutorial by Wand and Nott
[edit]
Reference notes
- devices
To see all active devices
dev.list()
To change device
X11()
postscript("file.eps", horizontal=FALSE, height=8, width=6, pointsize=10, onefile=FALSE) # file.ps could be used as well
pdf()
png()
jpeg()
Don't forget to terminate the device driver when done
dev.off()
- plot
plot(object, main="title",
xlab="x-label", xlim=c(xmin,xmax), # same applies for y
type="l", # l for lines, p for points, b for both
lab=c(nx_tics, ny_tics, xlabel_length),
log="xy" #indicates with axis to put on log scale
)
- put text on plot
xpos <- 10 ypos <- 10 text(xpos, ypos,expression(paste(the text in latex-like format goes here)))
- identify (identify points on your plot by clicking the button on mouse, right button returns a list of all points identified)
identify(object)
