Regex: negative look ahead

fcaw1 has asked for the wisdom of the Perl Monks concerning the following question:

Hi There I was wondering if you could advise why the following look ahead regex is not working properly. I would not expect the string to match because of the negative look ahead: CODE:

#!/app/perl-5.8/bin/perl

my $string = "ERROR blah license will expire";
print `/app/perl-5.8/bin/perl -version` . "\n\n";
if ( $string =~ /(ERROR.*?(?!(Error processing Cancel Execution|licens
+e will expire)))/ ){
  print "MATCHED\n";
} else {
  print "NOT MATCHED\n";
}
[download]

RESULT:

$ ./regex.pl

This is perl, v5.8.3 built for sun4-solaris

Copyright 1987-2003, Larry Wall

Perl may be copied only under the terms of either the Artistic License
+ or the
GNU General Public License, which may be found in the Perl 5 source ki
+t.

Complete documentation for Perl, including FAQ lists, should be found 
+on
this system using `man perl' or `perldoc perl'.  If you have access to
+ the
Internet, point your browser at http://www.perl.com/, the Perl Home Pa
+ge.



MATCHED
[download]

Many thanks for your help! Andy

Comment on Regex: negative look ahead Select or Download Code

Replies are listed 'Best First'.
Re: Regex: negative look ahead by JavaFan (Canon) on Apr 10, 2012 at 12:49 UTC
It seems you want to write a pattern that says "match PAT1, and there should not be PAT2 somewhere after it". You can write that as: `/PAT1(?!.PAT2)/s` [download] or, in your case, as `$string =~ /^ERROR(?!.\b(?:Error processing Cancel Execution\|license +will expire)\b)/ms` [download]	[reply] [d/l] [select]
Re: Regex: negative look ahead by halfcountplus (Hermit) on Apr 10, 2012 at 11:42 UTC
If the goal here is to do things with strings which do not match a pattern, don't try to use a negative look-ahead -- just negate the test: `my @strings = ( "ERROR blah license will expire", "ERROR blah something else", "ERROR blah Error processing Cancel Execution", ); foreach (@strings) { if (!($_ =~ /^ERROR.+?(Error processing Cancel Execution\|license w +ill expire)$/)) { print "Do something with '$_'\n"; } }` [download]	[reply] [d/l]
Re^2: Regex: negative look ahead by johngg (Canon) on Apr 10, 2012 at 12:43 UTC
Note that `if ( ! ( $string =~ /pattern/ ) ) {` [download] can be written `if ( $string !~ /pattern/ ) {` [download] I hope this is of interest. Cheers, JohnGG	[reply] [d/l] [select]
Re^2: Regex: negative look ahead by fcaw1 (Initiate) on Apr 10, 2012 at 12:32 UTC
Hi there I am writing a script with tails a log file and for each line runs multiple regular expressions against it (which are contained in a config file). Below are some examples of the log lines ERROR 2012-04-02 05:00:18 Error processing Cancel Execution '2012-CSNY +1CMR-1-20120402-1247 52938803o122h090000507' with LHDRSeqId 'CSNY1CMR +1-1333315812780-2759'. [2012-04-02 07:04:27,275][ERROR][101574:Company:japan][analytics.bond. +BondAlibImpl] ALIB license will expire in29 days. Please upgrade lice +nse file under /app/license 2012-04-02 05:00:18,810[ERROR] limitCheck: error while checking instru +ction xxx.core.biz.order.instruction.EquityOrderSubmitInstruction(tsI +d=20120402-090237887-340-509,order=20120402-090237880-767-508:TS) [download] With my original regular expression I would want to ignore the first two lines but match the third line. Thanks!	[reply] [d/l]
Re^3: Regex: negative look ahead by brx (Pilgrim) on Apr 10, 2012 at 14:37 UTC
It could be a good idea to split the big regex. A good example: if you want to match lines with the words 'red' and 'dog'. The first idea is often to write something like: `$line =~ /red.dog\|dog.red/` but there is a better way: `$line=~/dog/ and $line=~/red/` So, it's probably better to treat each exclusion with one regex: `next if /regex/;` while (my $line = <DATA>) { next if $line =~ /^ERROR.?Error processing Cancel Execution/; next if $line =~ /ERROR.?license will expire/; print "Do something with the line: $line"; } __DATA__ ERROR 2012-04-02 05:00:18 Error processing Cancel Execution '2012-CSNY +1CMR-1-20120402-1247 52938803o122h090000507' with LHDRSeqId 'CSNY1CMR +1-1333315812780-2759'. [2012-04-02 07:04:27,275][ERROR][101574:Company:japan][analytics.bond. +BondAlibImpl] ALIB license will expire in29 days. Please upgrade lice +nse file under /app/license 2012-04-02 05:00:18,810[ERROR] limitCheck: error while checking instru +ction xxx.core.biz.order.instruction.EquityOrderSubmitInstruction(tsI +d=20120402-090237887-340-509,order=20120402-090237880-767-508:TS) [download]	[reply] [d/l] [select]
Re: Regex: negative look ahead by Anonymous Monk on Apr 10, 2012 at 11:21 UTC
This is because you did not understood the regular expressions at the fundamental level. .*? zero chars -> true -> after the current position is not (?!....) -> true -> return true Use something like: `my $string = "ERROR blah license will expire"; if ($string =~ /^ERROR\b.+?(?:Error processing Cancel Execution\|licens +e will expire)\b/m) { print "NOT MATCHED\n"; } else { print "MATCHED\n"; }` [download] or: `if ($string =~ /^ERROR\b/m and $string !~ /\b(?:Error processing Cance +l Execution\|license will expire)\b/) { print "MATCHED\n"; } else { print "NOT MATCHED\n"; }` [download]	[reply] [d/l] [select]