Capturing groups where the ending is optional

cwm9 has asked for the wisdom of the Perl Monks concerning the following question:

I'm trying to split a line into two groups. The first group is a complete unknown, except that it will not contain a specific string. The second group starts with that same string and includes everything after it which may or may not be present. Example: suppose 'right' is the specific string in question. Here's an example of the desired input vs output:

left          \1=left \2=
right         \1=     \2=right
rightabc      \1=     \2=rightabc
leftright     \1=left \2=right
leftrightabc  \1=left \2=rightabc
[download]

Here's what I've tried so far. This version is overly greedy -- it won't give up eating group 2 into group 1.

s/(.*)(right)?/\1 <> \2/

echo "left"|perl -ne 'print if s/(.*)(right)?/\1 <> \2/'              
+  left <>
echo "right"|perl -ne 'print if s/(.*)(right)?/\1 <> \2/'             
+  right <>
echo "leftright"|perl -ne 'print if s/(.*)(right)?/\1 <> \2/'         
+  leftright <>
[download]

This version splits properly, but /2 is the same as /1 when there is no /2:

s/(.*(?=right.*)|(.*))/\

echo "left"|perl -ne 'print if s/(.*(?=right.*)|(.*))/\1 <> \2/'      
+  left <> left
echo "right"|perl -ne 'print if s/(.*(?=right.*)|(.*))/\1 <> \2/'     
+  <> right
echo "leftright"|perl -ne 'print if s/(.*(?=right.*)|(.*))/\1 <> \2/' 
+  left <> right
[download]

Comment on Capturing groups where the ending is optional Select or Download Code

Replies are listed 'Best First'.
Re: Capturing groups where the ending is optional by tybalt89 (Monsignor) on Jan 15, 2018 at 00:03 UTC
`#!/usr/bin/perl # http://perlmonks.org/?node_id=1207241 use strict; use warnings; my $specific = 'right'; while(<DATA>) { chomp; /(?\|(.?)($specific.)\|(.*)())/; printf "%-12s \\1=%4s \\2=%s\n", $_, $1, $2; } __DATA__ left right rightabc leftright leftrightabc` [download] Outputs: `left \1=left \2= right \1= \2=right rightabc \1= \2=rightabc leftright \1=left \2=right leftrightabc \1=left \2=rightabc` [download]	[reply] [d/l] [select]
Re: Capturing groups where the ending is optional by ikegami (Patriarch) on Jan 14, 2018 at 23:42 UTC
I think you're asking for `/ ^ ( (?: (?!\Q$key\E). )* ) ( .* ) /xs` [download] By the way, you shouldn't be using `\1` and `\2` in the substitution expression (as Perl told you through warnings you apparently ignored). `\1` and `\2` are regex atoms that instruct the regex engine to match first and second captured string respectively. It makes no sense to use them in the substitution expression. Use `$1` and `$2` instead.	[reply] [d/l] [select]
Re^2: Capturing groups where the ending is optional by cwm9 (Initiate) on Jan 14, 2018 at 23:49 UTC
Thank you. I'll examine this code. As for the $1 vs \1, there were no warnings --- I'm only using perl from the command line in a bash script. I'm pretty much a complete perl noob. Thank you for helping.	[reply]
Re^3: Capturing groups where the ending is optional by AnomalousMonk (Archbishop) on Jan 15, 2018 at 03:48 UTC
... there were no warnings --- I'm only using perl from the command line .... I'm pretty much a complete perl noob. Perl noobs in particular should always enable warnings and strictures. One way to do this from the command line is `perl -wMstrict -e "perl code here ..." whatever else ...` (double-quotes for Windoze, single-quotes otherwise). Give a man a fish: `<%-{-{-{-<`	[reply] [d/l] [select]
Re: Capturing groups where the ending is optional by AnomalousMonk (Archbishop) on Jan 15, 2018 at 00:30 UTC
Here, in the Test::More format that may, in future, avoid a great deal of confusion, is a possible solution based on ikegami's solution. Note that the `=>` (fat comma) in an expression like `[ 'leftrightabc' => 'left', 'rightabc' ],` is just a notational indulgence that is intended to be read as "produces" or "yields" and that is syntactically just a plain old `,` (comma) (update: because its LHS is fully quoted). c:\@Work\Perl\monks>perl -wMstrict -le "use Test::More 'no_plan'; use Test::NoWarnings; ;; use constant TEST_1 => ( [ 'left' => 'left', '' ], [ 'right' => '', 'right' ], [ 'rightabc' => '', 'rightabc' ], [ 'leftright' => 'left', 'right' ], [ 'leftrightabc' => 'left', 'rightabc' ], ); ;; my $specific_string = qr{ right }xms; ;; my $right_side = qr{ $specific_string .* \z }xms; my $left_side = qr{ (?! $specific_string) . }xms; ;; VECTOR: for my $ar_vector (TEST_1) { my ($string, @expected) = @$ar_vector; ;; is_deeply [ $string =~ m{ \A ($left_side) ($right_side?) }xms ], \@expected, qq{'$string' -> '$expected[0]' '$expected[1]'}; } ;; done_testing; " ok 1 - 'left' -> 'left' '' ok 2 - 'right' -> '' 'right' ok 3 - 'rightabc' -> '' 'rightabc' ok 4 - 'leftright' -> 'left' 'right' ok 5 - 'leftrightabc' -> 'left' 'rightabc' 1..5 ok 6 - no warnings 1..6 [download] Update:* Note also that `$specific_string` could be any `qr//` pattern. Give a man a fish: `<%-{-{-{-<`	[reply] [d/l] [select]