Listing all files matching a full-path pattern in R

问题

I am trying to obtain the list of files matching a full-path pattern. So far, I have used list.files() but it did not work.

Let's assume that we have the following directory organization:

results
   |- A
   |  |- data-1.csv
   |  |- data-2.csv
   |
   |- B
      |- data-1.csv
      |- data-2.csv

Then the following command:

list.files(pattern='data-.*\\.csv', recursive=TRUE)

will return all the files matching the pattern. This works, but the problem appears when using a full-path pattern. For instance, if I want to obtain all the CSV files from directory results/A, I could do:

list.files(pattern='results/A/data-.*\\.csv', recursive=TRUE)

This does not work, though. Somehow, it seems like R is not able to use a full-path pattern as a regular expression. In this case, the solution could be to just use results/A as the base path. But in more complex problems, that cannot be done. For instance, at some point we may want to match the subdirectories containing only characters:

list.files(pattern='results/[A-Z]+/data-.*\\.csv', recursive=TRUE)

Is it possible to do this in R?

UPDATE: After using ad hoc solutions for a while, I decided to stop typing the same again and again. So, I created a library for simplifying this task.

回答1:

First, note that you are not using regular expression patterns. Your first example should be:

list.files(pattern='data-.*\\.csv', recursive=TRUE)

Then, it seems the pattern matching inside list.files is applied to the file basenames (i.e., not including the directory path) so you could split the task into:

Find all files matching the basename only, return their full paths:

basename.matches <- list.files(pattern='data-.*\\.csv', recursive=TRUE,
                               full.names = TRUE)
basename.matches
# [1] "./results/A/data-1.csv" "./results/A/data-2.csv" "./results/B/data-1.csv"
# [4] "./results/B/data-2.csv"

Keep only those that match the expected directory(ies):

full.matches <- grep(pattern='^\\./results/A/', basename.matches, value = TRUE)
full.matches
# [1] "./results/A/data-1.csv" "./results/A/data-2.csv"

回答2:

You cannot do this with only list.files because it loops over each element in path and applies the regular expression to the files contained therein. But since the path argument to list.files can accept a vector, you can use that to solve your problem.

dirs <- grep("[A-Z]+$",list.dirs("results",recursive=FALSE),value=TRUE)
list.files(dirs, "data-.*\\.csv", recursive=TRUE, full.names=TRUE)

回答3:

I think there is an even simpler solution:

Sys.glob(file.path(results, "[A-Z]", "data-*.csv"))

回答4:

I will use

paths <- list.files(results, pattern= glob2rx("*data-*.csv$*"), full.names=T, recursive=T)

来源：https://stackoverflow.com/questions/10353540/listing-all-files-matching-a-full-path-pattern-in-r

标签

regex

path