Use regular expressions to find host name - regex

Question

so I have little problem, because I need to print host name which is bettwen "(?# )", for example:
Apr 17 23:39:02 test pure-ftpd: (?#researchscan425.eecs.umich.edu) [INFO] New connection from researchscan425.eecs.umich.edu
And I need to print "researchscan425.eecs.umich.edu".
I tried something like:
if(my ($test) = $linelist =~ /\b\(\?\#(\S*)/)
{
print "$test\n";
}
But it doesn't print me anything.

Answer 1

You can use this regex:
\(\?#(.*?)\)
researchscan425.eecs.umich.edu will be captured into Group 1.
See demo
Sample code:
my $linelist = 'Apr 17 23:39:02 test pure-ftpd: (?#researchscan425.eecs.umich.edu) [INFO] New connection from researchscan425.eecs.umich.edu';
if(my ($test) = $linelist =~ /\(\?#(.*?)\)/)
{
print "$test\n";
}

Answer 2

How about:
if(my ($test) = $linelist =~ /\(\?\#([^\s)]+)/)

Answer 3

You need to remove the \b which exists before (. Because there isn't a word boundary exists before ( (non-word character) and after space (non-word charcater).
my $linelist = 'Apr 17 23:39:02 test pure-ftpd: (?#researchscan425.eecs.umich.edu) [INFO] New connection from researchscan425.eecs.umich.edu';
if(my ($test) = $linelist =~ /\(\?\#([^)]*)/)
{
print "$test\n";
}

Answer 4

The problem here is the definition of \b.
It's "word boundary" - on regex101 that means:
(^\w|\w$|\W\w|\w\W)
Now, why this is causing you problems - ( is not a word character. So the transition from space to bracket doesn't trigger this pattern.
Switch your pattern to:
\s\(\?\#(\S+)
And it'll work. (Note - I've changed * to + because you probably want one or more, not zero or more).

Answer 5

It's amazing what you can do with logging tools or with perl as part of the logging service itself (c.f. Ubic), but even if you're just writing a "quick script" to parse logs for reporting (i.e. something you or someone else won't look at again for months or years) it helps to make them easy to maintain.
One approach to doing this is to process the lines of your log file lines with Regexp::Common. One advantage is that RX::Common matches practically "self document" what you are doing. For example, to match on specific "RFC compliant" definitions of what constitutes a "domain" using the $linelist you posted:
use Regexp::Common qw /net/;
if ( $line =~ /\?\#$RE{net}{domain}{-keep}/ ) { say $1 }
Then, later, if you need you can add other matches e.g "numeric" IPv4 or IPv6 addresses, assign them for use later in the script, etc. (Perl6::Form and IO::All used for demonstration purposes only - try them out!):
use IO::All ;
use Regexp::Common qw/net/;
use Perl6::Form;
my $purelog = io 'logfile.lines.txt' ;
sub _get_ftphost_names {
my #hosts = () ;
while ($_ = $purelog->getline) {
/\(\?\#$RE{net}{IPv6}{-sep => ":" }{-keep}/ ||
/\(\?\#$RE{net}{IPv4}{-keep}/ ||
/\(\?\#$RE{net}{domain}{-keep}/ and push #hosts , $1 ;
}
return \#hosts ;
}
sub _get_bytes_transfered {
... ;
}
my #host_list = _get_ftphost_names ;
print form
"{[[[[[[[[[[(30+)[[[[[[[[[[[[[}", #host_list ;
One of the great things about Regexp::Common (besides stealing regexp ideas from the source) is that it also makes it fairly easy to roll your own matches, You can use those to capture other parts of the file in an easily understandable way adding them piece by piece. Then, as what was supposed to be your four line script grows and transforms itself into a ITIL compliant corporate reporting tool, you and your career can advance apace :-)