TASdvlper has asked for the wisdom of the Perl Monks concerning the following question:

Hello all,

I was wondering is there a "quick/lightweight" perl in-line command to search the current directory, and all it's sub-directories, and look for a particular keyword (or words) in each file ?

I'm sure I can script, but I was hoping just to run something from the command line. Not that it's particularly better than a script.

Basically, I have a lot of code, in many diretories, and sometimes I need to look for a particular variable and sometimes I forgot what files contain those variable names. (Just one example of a use for this code).

Thanks all ....

  • Comment on Searching word(s) in multiple file and directories

Replies are listed 'Best First'.
Re: Searching word(s) in multiple file and directories
by Roger (Parson) on Dec 03, 2003 at 19:43 UTC
    You could use the grep utility if you are working on Unix platforms.

    If you are working under Windows, you could use the find utility.

    I have written a simple search.pl script in perl to search a directory tree for files containing a regex pattern. It's only tested under Windows.
    use strict;
use warnings;
use File::Find;
use Getopt::Long;
use Pod::Usage;
use Carp;

# Parse command line arguments and assign corresponding variables
GetOptions
(
    'r|rootdir=s'   => \( my $rootdir = "./" ),
    'p|pattern=s'   => \( my $pattern = undef ),
    'i'             => \( my $case_sensitive = 0 ),
    'v|verbose'     => \( my $verbose = 0 ),
);

unless ( defined $pattern )
{
    print <<USAGE
Usage: $0 [options]
Options:
 -r|--rootdir [dir]      Specify the top directory
 -p|--pattern [pattern]  Specify the pattern to look for (regex)
 -i                      Case sensitive search
 -v|--verbose            Print more info
USAGE
;
    exit(0);
}

$pattern = "(?i)" . $pattern if $case_sensitive;
print "looking under $rootdir for { /$pattern/ }\n" if $verbose;

find({ wanted => \&filefilter }, $rootdir);

sub filefilter
{
    return if /^\.+$/;
    open FILE, "<$File::Find::name"
         or carp "could not open file: $File::Find::name";
    my $file = do { local $/; <FILE> };
    close FILE;

    print ">> ", $File::Find::name, "\n"
        if $file =~ /$pattern/;
}

    [download]
    And the output -
    P:\Perl>perl search.pl -r ./ -p  "(Find|Dir)" -i
>> ./f1.pl
>> ./f15.pl
>> ./f2.pl
>> ./f48.pl
>> ./f6.pl
>> ./fun1.pl
>> ./Fun2.pl
>> ./result.txt
>> ./search.pl

    [download]
[reply]
[d/l]
[select]
Re: Searching word(s) in multiple file and directories
by flounder99 (Friar) on Dec 03, 2003 at 20:26 UTC
[reply]
[d/l]
Re: Searching word(s) in multiple file and directories
by Zed_Lopez (Chaplain) on Dec 03, 2003 at 20:29 UTC

    non-Perl unix approach:

    find . -exec fgrep variable_name {} /dev/null ;

    [download]

    Of course, you can substitute any directory for '.'; you need to surround fgrep's argument with single-quotes if it has shell-interpretable characters like '$'; you can use grep instead of fgrep to search by regexp instead of for plaintext.

    I have this in my .bashrc: 

    rfind() {
  find . -exec fgrep $1 {} /dev/null \;
}

    [download]
[reply]
[d/l]
[select]
      FWIW, some flavors of grep handle recursion on their own with the -r flag; check your local manpage for confirmation or just try it: 
        fgrep -r variable_name .

      [download]
[reply]
[d/l]
        Just what I was looking for. thanks.
[reply]
      Or to just print the filenames with the matching text: 
      find /path -type f -exec grep -l "StringToMatch"  "{}" \;

      [download]
      from there you can pipe on and actually do something with the output. Note that the -l is a lowercase L.


      -Waswas
[reply]
[d/l]

Back to Seekers of Perl Wisdom