perl: split filename.jpg into $prefix.$ext

ezekieldas · 09-29-2011, 10:18 AM

Howdy. Confused about this one. I have this loop that matches certain extensions but I can't seem to get the filename prefix out of it. Any suggestions are appreciated.

PHP Code:



#!/usr/bin/perl 
 
my $dir = "."; 
opendir( DIR, $dir ); 
my @files = readdir( DIR ); 
closedir( DIR ); 
 
sub get_rand_string { 
 my @chars=('a'..'z', 'A' .. 'Z', '0' .. '9'); 
 $rand_string = join("", @chars[ map { rand @chars } ( 1 .. 3 ) ]); 
 return $rand_string; 
} 
 
foreach my $file( @files ) { 
    if ( $file =~ /\w+\.(jpg|JPG|gif)$/ ) { 
 
#       how can I grab the prefix here? I want to split 
#       the filename into $prefix.$ext  
# 
#       As seen below I can get the $1 from the if statement 
#       above. 
# 
#    I've tried the following...  
# 
#       $orig_prefix = s/\.[^.]*$//; 
#       $orig_prefix =~ s/\.\w+$//; 
#       print "[$orig_prefix]\n"; 
#       $orig_prefix .= FOO; 
 
        my $ext = $1; 
 
        my $rand_string=&get_rand_string(); 
 
#     So ideally I'd like to place the prefix back on the filename  
#    with just the random string inserted...  
 
        my $filename_rand = sprintf( "$orig_prefix_$rand_string.$ext" ); 
#        my $filename_rand = sprintf( "$rand_string.$ext" ); 
 
        system( "mv $file $filename_rand" ); 
    } 
}

kbp · 09-29-2011, 06:21 PM

Been a little while since I've done perl but basically changing it like so -

Code:

if ( $file =~ /(w+).(jpg|JPG|gif)$/ ) {

..will give you 2 back references - $1 and $2, $1 will be the filename and $2 will be the extension

ezekieldas · 09-29-2011, 06:40 PM

Hey --thanks a bunch kbp! I still need to anchor (?) the w+ though:

if ( $file =~ /(\w+)\.(jpg|JPG|gif)$/ ) {

Just so I understand this better --do you, or does anyone, know what this is called technically? I iterate through elements of an array and if they match they're assigned $1, $2, etc? How would you communicate that in proper perl-speak?

kbp · 09-29-2011, 07:58 PM

The matching parts that are surrounded by '()' are called backreferences. You don't need to anchor the front unless you want to as you're specifically only matching words (\w).

They're called back references as you commonly use them in search/replace like:

Code:

's/(first_bit_I_want).*(second_bit_I_want).*(third_bit_I_want)/$1:$3:$2/'

From the example you can see that you can swap the ordering of the matching parts, add delimiters etc

anomie · 09-30-2011, 11:14 AM

Also, another approach with no regex. Use split():

Code:

my $filename = 'perky.jpg' ;

my @tokens = split(/\./, $filename) ;

print shift(@tokens) . "\n" ;

ezekieldas · 10-03-2011, 05:09 PM

Thanks again to all the replies here -- very helpful!