I want to cbind several files together using:
do.call("cbind",lapply(sample_list, FUN=function(files){read.table(files, header=TRUE, sep="\t", stringsAsFactors=FALSE)}))
However, my sample_list files (eg, 1c.QC.dat) are in different directories. But the directories follow the same pattern:
/home/Project1/Files/Sample_*/*.QC.dat
where * is the sample ID.
Is there a way to list these files easily?
Let's first select our Sample_*
directories.
main_dir <- '/home/Project1/Files/'
directories <- list.files(main_dir, pattern = '^Sample_')
directories <- Filter(function(x) file.info(file.path(main_dir, x))$isdir, directories)
We now have a character vector of directories beginning with Sample_
. Now we can read in our data.frames:
dfs <- lapply(directories, function(subdir) {
files <- list.files(path <- file.path(main_dir, subdir), pattern = '\\.QC\\.dat$')
subdfs <- lapply(files, function(filename)
read.table(file.path(path, filename), header=TRUE, sep="\t", stringsAsFactors=FALSE)
)
do.call(rbind, subdfs)
})
Finally, we bind them into one giant dataframe:
dfs <- do.call(rbind, dfs) # Notice we used the same trick twice
A shorter but cleverer option is to use the recursive = TRUE
parameter on list.files
:
dfs <- do.call(rbind, lapply(
list.files(path <- '/home/Project1/Files/',
pattern = '^Sample_.*\\.QC\\.dat$', recursive = TRUE),
function(filename)
read.table(file.path(path, filename), header=TRUE, sep="\t", stringsAsFactors=FALSE)
))