在单个命令行中执行grep和sed

Question

I have written a perl script with two arrays. My first array contains a list of names and my second array contains a list of names that I would like to replace the first names with.

This is a quick and dirty script I'm writing just to get something done so I do not have to do the manual work. It's not something I am focusing on looking "clean" or super efficient. For that reason, I'm making use of system a lot.

A sample directory that I am using would look like this:

/myDir/Names/John/ containing files John.log, John.txt, RandomFile.html

John.log

John is an engineer.
Jack and John are friends.

John.txt

John is 48 years old. 

where John is an element in my first array. I want to be able to replace John with, for example, Mary from my second array IN FILES THAT CONTAIN JOHN IN THE FILE NAME. I want the element of that array to change to the corresponding element number in the next array. Here is what I've come up with:

#!/usr/bin/perl
use strict;
use warnings;

my @firstnames = (John, Jack, Jill);
my @secondnames = (Mary, Nick, Joseph);

foreach my $old (@firstnames){
    foreach my $new (@secondnames){
         system("grep -rl $old /myDir/Names/$old/ | xargs sed -i 's/$old/$new/g' $old.* ");
        system("rename '$old' '$new' /myDir/Names/$old.*");
     }
}

Grep seems to work fine, Rename and Sed does not because of the '.*'. The issue is where I have $old.* . It appropriately substitutes $old with the correct element in the array but it looks for files called John.* , Jack.* , Jill.* instead of reading .* as any file extension. Clearly, I am not understanding how sed works entirely. I would like some help as to how I can do the sed command for all files containing $old wit ANY file extension, because there might be multiple files in the directories with same names and different extensions. The reason I specify the first name with any extension is because the file might contain something like RandomFile.html that I listed above and I only want files that contain $old before extension.

This is the output that I want, where the files names along with the words found in the text will be replaced with the corresponding element from the second array:

Mary.log

Mary is an engineer.
Jack and Mary are friends.

Mary.txt

Mary is 48 years old. 

Note: The words that are in the file, match the name of the file they are in, which match the name of the directory they are in. That is the pattern of these Directories and their files.

Please let me know if anything is unclear and I will edit my post.

Answer 1

So the issue here is that you're piping a list of filenames into xargs already which get passed to sed. You cannot have sed take the intersection of all filenames that have the desired words in them AND the contents of those files as well. You will need to first find all the files with the desired contents, filter out the filenames you do not care about, THEN feed that list into xargs sed. Try:

system("grep -rl $old /myDir/Names/$old/ | grep $old | xargs sed -i 's/$old/$new/g'");

Answer 2

I was having issues with grep that I was not able to entirely trace down so I used find instead which seems to do the trick. Here is what I'm using now that works just fine:

find ./$old -name "$old.*" -exec gsed -i -e "s/$old/$new/Ig" {} +

Another important issue to note that is not particularly related to the problem in the question is that I am incorrectly using the foreach loop. In order to match each element in the arrays, you have to iterate through the size of the array.

#!/usr/bin/perl
use strict;
use warnings;

my $i = scalar @firstnames;
for ($i=0; $i < @firstnames; $i++){
.......
 }

Answer 3

.Hi sfr. This will do the job although there is no error checking.

#!perl

use strict;
use diagnostics;
use File::Glob "bsd_glob";
use List::MoreUtils "mesh";
use Data::Dumper "Dumper";
our $basedir = "/myDir/Names";
our @oldnames = qw/John Jack Jill/;
our @newnames = qw/Mary Nick Joseph/;
our %namemap = mesh @oldnames, @newnames;
printf "oldnames:%s\nnewnames: %s\nnamemap:%s\n\n", map Dumper($_), \@oldnames, \@newnames, \%namemap;

foreach my $searchname (keys %namemap) {
    print "searchname: $searchname, glob: $basedir/$searchname/*$searchname*";
    my @files = bsd_glob "$basedir/$searchname/*$searchname*";
    printf "files %s\n", Dumper \@files;

    next unless @files;
    local $/;
    foreach my $file (@files) {
        print "on file $file\n";
        open my $F, "+<", $file or die "open $file: $!";
        my $text = readline $F;
        print "changing all $searchname to $namemap{$searchname}\n";
        if (my $c = $text =~ s/\b$searchname\b/$namemap{$searchname}/g) {
            print "replaced $c occurrences\n";
            seek $F, 0, 0 or die "seek $file: $!";
            print $F $text or die "print $file: $!";
            print "wrote changes to file\n";
            close $F or die "closing $file: $!";
        }
        else { print "file did not contain any $searchname\n"; }
        (my $newfile = $file) =~ s/\b$searchname\b/$namemap{$searchname}/g;
        print "renaming $file to $newfile\n";
        rename $file, $newfile or die "rename: $!";
    }
}

LA

EDIT: code updated