How can I extract a file path from a Perl string?

Question

I want to find the file name with fullpath from this string

"[exec] /snschome/sns/ccyp/mb_ccsns/bb/cbil5/v85_8/amdocs/iamcust/bil/cl/business/handlers/ClReportsHandler.java:233: cannot resolve symbol"

I want to extract

/snschome/sns/ccyp/mb_ccsns/bb/cbil5/v85_8/amdocs/iamcust/bil/cl/business/handlers/ClReportsHandler.java

and I am trying this in Perl

    $_=$string_from_to_match
    my @extract_file=split(/.*(\/.*)\:.*/);
    print $extract_file[1],"\n";`

but I am getting this output:

/ClReportsHandler.java:233:

It is matching the last / and the last :. How can I change it to first / and first :?

Answer 1

Why are you using split() if you want to match?

if ($subject =~ m!(/[^:]+):!) {
    $result = $1;
} else {
    $result = "";
}

should do it.

Explanation:

!: alternative regex delimiter (since I need the / inside)

(: start capturing group

/: match a /. This will always be the first / of the string.

[^:]+: match 1 or more characters except :

): end capturing group

:: match a :

!: regex delimiter

Answer 2

If your file name doesn't contain spaces then you can create simple regexp to match all parts:

my ($file, $line, $msg) = ( $string_from_to_match =~ m{(\S+):([^:]+):([^:]+)} );

I've used:

• \S+ to match 1 or more non-space symbols
• [^:]+ to match 1 or more not : symbols

If you want spaces in path, then the best way is to remove starting [exec] part and split by ::

$string_from_to_match =~ s{^\[exec\]\s+}{};
my ($file, $line, $msg) = split(/:/, $string_from_to_match, 3);

Answer 3

This is a case where “tacking and stretching” is useful. You know that [exec] followed by whitespace is on the left and colon followed by a line number is on the right. You want what's in between:

#! /usr/bin/perl

use warnings;
use strict;

$_ = "[exec] /snschome/sns/ccyp/mb_ccsns/bb/cbil5/v85_8/amdocs/iamcust/bil/cl/business/handlers/ClReportsHandler.java:233: cannot resolve symbol";

if (/\[exec\]\s*(.+?):\d+/) {
  print $1, "\n";
}

Answer 4

You could try :

#!/usr/bin/perl
use strict;
use warnings;
use 5.10.1;

my $str = "[exec] /snschome/sns/ccyp/mb_ccsns/bb/cbil5/v85_8/amdocs/iamcust/bil/cl/business/handlers/ClReportsHandler.java:233: cannot resolve symbol";
$str =~ s!^[^/]*(/[^:]*):.*$!$1!;
say $str;

Ouput: /snschome/sns/ccyp/mb_ccsns/bb/cbil5/v85_8/amdocs/iamcust/bil/cl/business/handlers/ClReportsHandler.java