How to iterate over file names in a R script?

Question

A very simple R script adds a ascending row index to a textfile "0.txt" and also adds a header "time" to it. After that the data is written to a file "0-edit.txt"

data<-read.table("0.txt", header=TRUE,sep=",",row.names= NULL);
colnames(data)[1] = "time"
write.table(data,quote=FALSE,sep=", ","0-edit.txt");

Assume i have 4 files called 0.txt, 1.txt, 2.txt, ...in the same folder, how can I use a counter (or something else) to iterate over those file names in my script?

score 15 · Accepted Answer · answered Jan 31 '11 at 17:45

15

for(i in 0:3) {
  infile <- paste(i,".txt",sep="")
  outfile <- paste(i,"-edit.txt",sep="")

  data <- read.table(infile,header=TRUE,sep=",",row.names=NULL)
  colnames(data)[1] = "time"
  write.table(data,quote=FALSE,sep=", ",outfile)
}

answered Jan 31 '11 at 17:45

Joshua Ulrich

173,410
32
338
418

Thank you very much, the paste function escaped me somehow – Martin H Jan 31 '11 at 18:00
1

@Martin: You're welcome. You may also find `?file.path` handy. – Joshua Ulrich Jan 31 '11 at 18:06

score 15 · Answer 2 · answered Feb 01 '11 at 06:58

Here is a solution without a loop using lapply:

infiles <- dir(pattern='\\.txt$')

change.files <- function(file){
  data <- read.table(file, header=TRUE, sep=",", row.names=NULL)
  colnames(data)[1] = "time"
  write.table(data, quote=FALSE, sep=", ", sub("\\.txt$","-edit.txt", file))
}

lapply(infiles , change.files)

For me the real advantage of this approach is that you can easily run it in parallel using mclapply (from multicore package) instead of lapply. Or parLapply from snow. Also to me it looks nicer.

This just came in handy for another task and I have to agree, it is elegant — Martin H, Apr 17 '12 at 13:04

score 7 · Answer 3 · answered Feb 01 '11 at 16:40

Try this:

files <- list.files(path="", pattern=".txt", all.files=T, full.names=T)
for (file in files) {
## do stuff
}

You can use regular expressions for the pattern matching, so if you have lots of text files but only want the ones with numeric names, use "[0-9].txt" or "[0-3].txt".

score 3 · Answer 4 · answered Jan 31 '11 at 20:47

More generally, you can use dir() to get the files in a given directory, and use select to limit it to .txt files.

file.dir <- "/path/to/files"
for(infile in dir(file.dir, pattern="\\.txt$")) {
  outfile <- gsub("\\.txt$","-edit\\.txt", infile)

  data <- read.table(infile,header=TRUE,sep=",",row.names=NULL)
  colnames(data)[1] = "time"
  write.table(data,quote=FALSE,sep=", ",outfile)
}

How to iterate over file names in a R script?

4 Answers4

Linked