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

I need to search for a specific piece of text in all files in a directory. Since I also need to do this recursively I want to use File::Find.

I can return file names easily. What I need to do, though, is to search for the pattern then only print out those files which have the pattern in it. I don't want to do anything to the files; I just want a list returned.

I know it's simple but for some reason I have thrown up a brick wall.

I'm assuming (I know that can be dangerous) that I would need to do an open(FH, $file) on each file returned then search for the pattern then.

Am I even on the right track?

Thanks.

Replies are listed 'Best First'.
Re: text search in a file
by shemp (Deacon) on Aug 10, 2005 at 21:55 UTC
    Update:I had the args to grep backwards
    I dont know what else you'll need to do with the info, since you said that you're going to print out the filenames, if you're on a *nix system you could just:

    grep -Rl <basedir> <string to macth> 

    grep -Rl <string to macth> <basedir>

    [download]
    The 'R' switch is grep recursive, and the 'l' switch is to have grep only output the filenames with matches, instead of the full lines.
    I use the most powerful debugger available: print!
[reply]
[d/l]
      other way around (man grep for the switch details): 
      grep [options] PATTERN [FILE...]

      [download]
[reply]
[d/l]
[select]
      Unfortunately, not all greps have a recursive flag, such as /usr/bin/grep on Solaris. :( I'm forced to create my own recursive grep like this: 
      grep -li search_term * */* */*/* */*/*/* (etc.)

      [download]
      But I agree that a recursive grep is the best solution for the OP.
[reply]
[d/l]

        Nah - much easier to do something like: 

        grep -li search_term `find . -type f`

        [download]
        Assuming there aren't so many files that you exhaust the command shell line length. Or you can check out xargs, or even -exec option on find.

        TMTOWTDI - even in shell. :-)

[reply]
[d/l]

        Or if you use the most spiffy zsh you can do this: 

        grep -li search **/*(.)

        [download]

        Which will recursively search only files. But again that may bump into a number of arguments limitation (in which case you'd use print -N **/*(.) | xargs grep -li search or the like).

        --
        We're looking for people in ATL

[reply]
[d/l]
[select]

      Sheesh. *cough*

      Boy is *my* face red.

      I was so caught up in using perl I completely forgot about grep.
      That does exactly what I need.

      Thanks muchly.

[reply]
Re: text search in a file
by sgifford (Prior) on Aug 10, 2005 at 22:23 UTC
    A straightforward way to do this is write a wanted sub to search for the pattern, and if it matches append the current filename to an array.

    Here's an example: 

    #!/usr/bin/perl

use warnings;
use strict;

use File::Find;
use FileHandle;

# Pattern is first argument;
# remaining arguments are file or directory names.
# Default directory is '.'

my $pat = shift;
my @results;

# Do the search
find(patternfinder(qr/$pat/,\@results),
     @ARGV?@ARGV:'.');

# Print the results
print join("\n",@results),"\n";

sub patternfinder
{
  # Generate a sub to search for a pattern and save the results
  my($pattern,$result) = @_;

  # Anonymous sub in a closure
  return sub {
    my $fh = new FileHandle $_, "r"
        or return warn "Couldn't open '$File::Find::name': $!\n";

    # Search for pattern, and add to array if it matches.
    while (<$fh>) {
      if (/$pattern/) {
        push(@$result,$File::Find::name);
        return;
      }
    }
    
    close($fh)
        or return warn "Couldn't close '$File::Find::name': $!\n";
  }
}

    [download]
[reply]
[d/l]
[select]
Re: text search in a file
by gellyfish (Monsignor) on Aug 10, 2005 at 21:53 UTC

    A suggestion would be to look at the code of the NMS Simple Search which does exactly this.

    /J\

[reply]

Back to Seekers of Perl Wisdom