These functions provide a low-level interface to the computer's file system.
file.create(..., showWarnings = TRUE) file.exists(...) file.remove(...) file.rename(from, to) file.append(file1, file2) file.copy(from, to, overwrite = recursive, recursive = FALSE, copy.mode = TRUE, copy.date = FALSE) file.symlink(from, to) file.link(from, to)
character vectors, containing file names or paths.
character vectors, containing file names or paths.
logical; should existing destination files be overwritten?
logical; should the warnings on failure be shown?
logical: should file permission bits be copied where possible?
logical: should file dates be preserved where
... arguments are concatenated to form one character
string: you can specify the files separately or as one vector.
All of these functions expand path names: see
file.create creates files with the given names if they do not
already exist and truncates them if they do. They are created with
the maximal read/write permissions allowed by the
‘umask’ setting (where relevant). By default a warning
is given (with the reason) if the operation fails.
file.exists returns a logical vector indicating whether the
files named by its argument exist. (Here ‘exists’ is in the
sense of the system's
stat call: a file will be reported as
existing only if you have the permissions needed by
Existence can also be checked by
might use different permissions and so obtain a different result.
Note that the existence of a file does not imply that it is readable:
for that use
file.access.) What constitutes a
‘file’ is system-dependent, but should include directories.
(However, directory names must not include a trailing backslash or
slash on Windows.) Note that if the file is a symbolic link on a
Unix-alike, the result indicates if the link points to an actual file,
not just if the link exists.
Lastly, note the different function
checks for existence of R objects.
file.remove attempts to remove the files named in its argument.
On most Unix platforms ‘file’ includes empty
directories, symbolic links, fifos and sockets. On Windows,
‘file’ means a regular file and not, say, an empty directory.
file.rename attempts to rename files (and
to must be of the same length). Where file permissions allow
this will overwrite an existing element of
to. This is subject
to the limitations of the OS's corresponding system call (see
man 2 rename on a Unix-alike): in particular
in the interpretation of ‘file’: most platforms will not rename
files from one file system to another. NB: This means that
renaming a file from a temporary directory to the user's filespace or
during package installation will often fail. (On Windows,
file.rename can rename files but not directories across
volumes.) On platforms which allow directories to be renamed,
typically neither or both of
to must a
directory, and if
to exists it must be an empty directory.
file.append attempts to append the files named by its
second argument to those named by its first. The R subscript
recycling rule is used to align names given in vectors
of different lengths.
file.copy works in a similar way to
file.append but with
the arguments in the natural order for copying. Copying to existing
destination files is skipped unless
overwrite = TRUE. The
to argument can specify a single existing directory. If
copy.mode = TRUE file read/write/execute permissions are copied
where possible, restricted by ‘umask’. (On Windows this
applies only to files.) Other security attributes such as ACLs are not
copied. On a POSIX filesystem the targets of symbolic links will be
copied rather than the links themselves, and hard links are copied
copy.date = TRUE may or may not copy the
timestamp exactly (for example, fractional seconds may be omitted),
but is more likely to do so as from R 3.4.0.
file.link make symbolic and hard links
on those file systems which support them. For
to argument can specify a single existing directory. (Unix and
macOS native filesystems support both. Windows has hard links to
files on NTFS file systems and concepts related to symbolic links on
recent versions: see the section below on the Windows version of this
help page. What happens on a FAT or SMB-mounted file system is OS-specific.)
File arguments with a marked encoding (see
if possible translated to the native encoding, except on Windows where
Unicode file operations are used (so marking as UTF-8 can be used to
access file paths not in the native encoding on suitable file
These functions return a logical vector indicating which operation succeeded for each of the files attempted. Using a missing value for a file or path name will always be regarded as a failure.
showWarnings = TRUE,
file.create will give a warning
for an unexpected failure.
Case-insensitive file systems are the norm on Windows and macOS, but can be found on all OSes (for example a FAT-formatted USB drive is probably case-insensitive).
These functions will most likely match existing files regardless of case on such file systems: however this is an OS function and it is possible that file names might be mapped to upper or lower case.
Always check the return value of these functions when used in package
code. This is especially important for
file.rename, which has
OS-specific restrictions (and note that the session temporary
directory is commonly on a different file system from the working
directory): it is only portable to use
file.rename to change
file name(s) within a single directory.
Ross Ihaka, Brian Ripley
Sys.glob to expand wildcards in file specifications.
Sys.readlink (for ‘symlink’s).
cat("file A\n", file = "A") cat("file B\n", file = "B") file.append("A", "B") file.create("A") # (trashing previous) file.append("A", rep("B", 10)) if(interactive()) file.show("A") # -> the 10 lines from 'B' file.copy("A", "C") dir.create("tmp") file.copy(c("A", "B"), "tmp") list.files("tmp") # -> "A" and "B" setwd("tmp") file.remove("A") # the tmp/A file file.symlink(file.path("..", c("A", "B")), ".") # |--> (TRUE,FALSE) : ok for A but not B as it exists already setwd("..") unlink("tmp", recursive = TRUE) file.remove("A", "B", "C")
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.