Splitting a Filename into Its Component Parts

Problem

You want to extract a filename, its enclosing directory, or the extension(s) from a string that contains a full pathname.

Solution

Use routines from the standard File::Basename module.

use File::Basename; $base = basename($path); $dir = dirname($path); ($base, $dir, $ext) = fileparse($path);

Discussion

The standard File::Basename module contains routines to split up a filename. dirname and basename supply the directory and filename portions respectively:

$path = '/usr/lib/libc.a'; $file = basename($path); $dir = dirname($path);
print "dir is $dir, file is $file\n"; # dir is /usr/lib, file is libc.a

The fileparse function can be used to extract the extension. To do so, pass fileparse the path to decipher and a regular expression that matches the extension. You must give fileparse this pattern because an extension isn't necessarily dot-separated. Consider ".tar.gz"--is the extension ".tar", ".gz", or ".tar.gz"? By specifying the pattern, you control which of these you get.

$path = '/usr/lib/libc.a'; ($name,$dir,$ext) = fileparse($path,'\..*');
print "dir is $dir, name is $name, extension is $ext\n"; # dir is /usr/lib/, name is libc, extension is .a

's normal conventions for directory separators by looking at the $^O variable, which holds a string identifying the system you're running on. That value was determined when Perl was built and installed. You can change the default by calling the fileparse_set_fstype routine. This alters the behavior of subsequent calls to the File::Basename functions:

fileparse_set_fstype("MacOS"); $path = "Hard%20Drive:System%20Folder:README.txt"; ($name,$dir,$ext) = fileparse($path,'\..*');
print "dir is $dir, name is $name, extension is $ext\n"; # dir is Hard%20Drive:System%20Folder, name is README, extension is .txt

To pull out just the extension, you might use this:

sub extension {
 my $path = shift; my $ext = (fileparse($path,'\..*'))[2]; $ext =~ s/^\.//;
return $ext;
}

When called on a file like source.c.bak, this returns an extension of "c.bak", not just "bak". If you wanted just ".bak" returned, use '\..*?' as the second argument to fileparse.

When passed a pathname with a trailing directory separator, such as lib/, fileparse considers the directory name to be "lib/", whereas dirname considers it to be ".".