regexp look behind?

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

Hi All, How can I print a "\n" in between two different strings, cache and DISK.* where they appear twice.

__INPUT__
cache,23778,0.0,0,0.0,0,0.0,0.0,0.00
cache,205,0.0,0,0.0,0,0.0,0.0,0.00
DISK-UTIL-PEAK-HIGH:,54.3,DISK-UTIL:,2.5
PROC-Name,PID,PYS-IO-RATE,READs,READ-RATE,WRITEs
cache,26094,0.9,1,0.9,0,0.0,0.0,0.00
cache,16046,0.0,0,0.0,0,0.0,0.0,0.00
cache,3848,0.0,0,0.0,0,0.0,0.0,0.00
cache,23778,0.0,0,0.0,0,0.0,0.0,0.00
cache,205,0.0,0,0.0,0,0.0,0.0,0.00
DISK-UTIL-PEAK-HIGH:,54.3,DISK-UTIL:,3.4  ### last line ###
__END__

use strict;
use warnings;

my $dstattmp = qq(/usr/local/log/disk.gpm.statlog.tmp);
my $dstatlog = qq(/usr/local/log/disk.gpm.statlog.csv);

open( my $DT, "+>>", $dstatlog ) || die "$dstatlog was not opened $!";
open( my $TD, "+>",  $dstattmp ) || die "$dstattmp was not opened $!";

while (<$DT>) {
    next if ( $. == 1);
    if ( /cache|disk-.*/i ) {
        push my @tmp, split;
        print $TD join (",", @tmp),"\n";
    }

    if ( /^(\D.*)(?=\d*)(?=disk-util-peak.*)/i )
        print $TD "YES\n";
    }
    ##if ( /^(cache.*)(?=\d*)(?=disk-util-peak.*)/i )
    ##if ( /^(cache.*)(?=\d*)/ .. /disk-util-peak.*/igs )  
} 

thank you!
[download]

Comment on regexp look behind? Download Code

Replies are listed 'Best First'.
Re: regexp look behind? by raybies (Chaplain) on Nov 30, 2010 at 16:03 UTC
keep track of your last line's match. If you'll wrap your regex in parenthesis like so `if (/(cache\|disk-)/i)`, you can assign the match to a variable, like so `$lastmatch = $1;`, then each pass through your loop, you then can `print "\n" if ($lastmatch ne $1 and "disk-" eq lc $1);`	[reply] [d/l] [select]
Re^2: regexp look behind? by dbs (Sexton) on Nov 30, 2010 at 19:24 UTC
thanks but no cigar! I would like the output to be `cache,23778,0.0,0,0.0,0,0.0,0.0,0.00 cache,205,0.0,0,0.0,0,0.0,0.0,0.00 "newline here" DISK-UTIL-PEAK-HIGH:,54.3,DISK-UTIL:,2.5 "newline here" PROC-Name,PID,PYS-IO-RATE,READs,READ-RATE,WRITEs cache,26094,0.9,1,0.9,0,0.0,0.0,0.00 cache,16046,0.0,0,0.0,0,0.0,0.0,0.00 cache,3848,0.0,0,0.0,0,0.0,0.0,0.00 cache,23778,0.0,0,0.0,0,0.0,0.0,0.00 cache,205,0.0,0,0.0,0,0.0,0.0,0.00 "newline here" DISK-UTIL-PEAK-HIGH:,54.3,DISK-UTIL:,3 ### last line ###` [download]	[reply] [d/l]
Re^3: regexp look behind? by AnomalousMonk (Archbishop) on Dec 01, 2010 at 01:25 UTC
No, thank you for finally clarifying just what you want the output to be!	[reply]
Re^4: regexp look behind? by raybies (Chaplain) on Dec 01, 2010 at 13:21 UTC
Re: regexp look behind? by jffry (Hermit) on Dec 01, 2010 at 00:47 UTC
This is kind of a kludge, and it ain't using fancy REs, but I think it does what you want. `#!/usr/bin/perl -pw use strict; use warnings; our ($prev1, $prev2); if (/DISK.DISK/) { if ($prev1 =~ /cache/ && $prev2 =~ /cache/) { print "\n"; } } if ($prev1 =~ /DISK.DISK/) { print "\n"; } $prev2 = $prev1; $prev1 = $_;` [download] Example execution with sample data. me@mybox:~/sandbox $ ./2.pl input.txt Use of uninitialized value in pattern match (m//) at ./2.pl line 14, < +> line 1. cache,23778,0.0,0,0.0,0,0.0,0.0,0.00 cache,205,0.0,0,0.0,0,0.0,0.0,0.00 DISK-UTIL-PEAK-HIGH:,54.3,DISK-UTIL:,2.5 PROC-Name,PID,PYS-IO-RATE,READs,READ-RATE,WRITEs cache,26094,0.9,1,0.9,0,0.0,0.0,0.00 cache,16046,0.0,0,0.0,0,0.0,0.0,0.00 cache,3848,0.0,0,0.0,0,0.0,0.0,0.00 cache,23778,0.0,0,0.0,0,0.0,0.0,0.00 cache,205,0.0,0,0.0,0,0.0,0.0,0.00 DISK-UTIL-PEAK-HIGH:,54.3,DISK-UTIL:,3.4 ### last line ### me@mybox:~/sandbox $ cat input.txt cache,23778,0.0,0,0.0,0,0.0,0.0,0.00 cache,205,0.0,0,0.0,0,0.0,0.0,0.00 DISK-UTIL-PEAK-HIGH:,54.3,DISK-UTIL:,2.5 PROC-Name,PID,PYS-IO-RATE,READs,READ-RATE,WRITEs cache,26094,0.9,1,0.9,0,0.0,0.0,0.00 cache,16046,0.0,0,0.0,0,0.0,0.0,0.00 cache,3848,0.0,0,0.0,0,0.0,0.0,0.00 cache,23778,0.0,0,0.0,0,0.0,0.0,0.00 cache,205,0.0,0,0.0,0,0.0,0.0,0.00 DISK-UTIL-PEAK-HIGH:,54.3,DISK-UTIL:,3.4 ### last line ### [download] FYI, what I think you're asking for is: if "cache" is on 2 consecutive lines followed by "DISK" twice on the 3rd consecutive line, then insert a newline before and after the "DISK" line except the last newline at the end.	[reply] [d/l] [select]