Re: Regex lookahead, lookbehind

Look around assertions are needed when you need overlapping matches, they are zero-width. Your case seems matchable without them, too:

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

while (<DATA>) {
    print "$. ok\n" if /^foo      # "foo"
                     _            # followed by "underscore"
                     (?:bar|test) # followed by "bar|test" (bar or tes
+t)
                     _            # followed by "underscore"

                     ((?:\w+\.)+) # followed by alhpnum string with
                                  # dot at end (like foo10.1.1.1)

                     \1?          # followed by one or zero occurences
                                  # of same alhpnum string (like
                                  # foo10.1.1.1) followed by dot

                     \w{1,10}     # followed by 1-10 characters,
                                  # containing at least one digit and
                                  # one letter and underscore (e.g.
                                  # TEST_test, TEST_test2,
                                  # TEST1_test2)

                     \.txt$       # ends with .txt
                    /x;
}

__DATA__
foo_bar_foo10.1.1.1.TEST.txt
foo_test_foo10.1.1.1.foo10.1.1.1.TEST_test.txt
[download]

To test the condition "at least one digit and one letter and underscore", I'd probably capture the group and test these conditions independently:

/\d/ && /[[:alpha:]]/ && /_/
[download]

($q=q:Sq=~/;[c](.)(.)/;chr(-||-|5+lengthSq)`"S|oS2"`map{chr |+ord
}map{substrSq`S_+|`|}3E|-|`7**2-3:)=~y+S|`+$1,++print+eval$q,q,a,
[download]

Comment on Re: Regex lookahead, lookbehind Select or Download Code

Replies are listed 'Best First'.
Re^2: Regex lookahead, lookbehind by rahulme81 (Sexton) on Feb 23, 2017 at 09:25 UTC
Thank you very much choroba. This works !!! One last thing I need to extend to below match also `foo_bar_foo10.1.1.1.TEST.txt foo_test_foo10.1.1.1.foo10.1.1.1.TEST_test.txt foo_test_foo10.1.1.1.foo10.1.1.1.txt foo_test_foo10.1.1.1.foo10.1.1.1.TEST-1.txt` [download] Third string is not a correct match but this is displayed as "Ok" whereas fourth doesn't seems ok but I want to match this also for "foo_test_foo10.1.1.1.foo10.1.1.1.TEST-1.txt" I modified the expression as (\w\-{1,10}) and that works ... Is this correct way ? trying to achieve the third format within the same format - as Nok Output for now 1 ok 2 ok 3 ok 4 ok Does this approach allows to use variable inside the regex format ? Can I declare like my $var = foo10.1.1.1 and use this in regex format ? Yes it does ... found answer for $var declaration I replaced `((?:\w+\.)+)` as `((?:($var)+\.)+)` ==> Is this correct too Last thing I am not able to get 3 is not correct .... I want to filter this out OR foo_test_foo10.1.1.1.foo10.1.1.1.txt - How to match only this in a separate regex	[reply] [d/l] [select]
Re^3: Regex lookahead, lookbehind by choroba (Cardinal) on Feb 23, 2017 at 10:04 UTC
I don't understand your requirements. It now seems to me none of the examples is correct, as TEST and TEST_test don't contain at least one digit, and TEST-1 doesn't contain at least one underscore. Changing the dash to underscore in the last makes it pass: #!/usr/bin/perl use warnings; use strict; while (<DATA>) { if (my ($alnum, $constrained) = /^foo # "foo" _ # followed by "underscore" (?:bar\|test) # followed by "bar\|test" (bar or test) _ # followed by "underscore" ((?:\w+\.)+?) # followed by alhpnum string with # dot at end (like foo10.1.1.1) \1? # followed by one or zero occurences # of same alhpnum string (like # foo10.1.1.1) followed by dot (\w{1,10}) # followed by 1-10 characters, # containing at least one digit and # one letter and underscore (e.g. # TEST_test, TEST_test2, # TEST1_test2) \.txt$ # ends with .txt /x) { print "<$alnum \| $constrained> $. ok\n" if 3 == grep $constrained =~ $_, qr/_/, qr/\d/, qr/[[:alpha:]]/; } } __DATA__ foo_bar_foo10.1.1.1.TEST.txt foo_test_foo10.1.1.1.foo10.1.1.1.TEST_test.txt foo_test_foo10.1.1.1.foo10.1.1.1.txt foo_test_foo10.1.1.1.foo10.1.1.1.TEST-1.txt foo_test_foo10.1.1.1.foo10.1.1.1.TEST_1.txt [download] Note that I changed the quantifier in the first capturing group to frugal to avoid matching both occurrences of $alnum as one without repetition. Nevertheless, in the third case, the whole substring `foo10.1.1.1.foo10.1.1.` corresponds to $alnum, it's not repeated, and $constrained is just `1` . ($q=q:Sq=~/;[c](.)(.)/;chr(-\|\|-\|5+lengthSq)`"S\|oS2"`map{chr \|+ord }map{substrSq`S_+\|`\|}3E\|-\|`7**2-3:)=~y+S\|`+$1,++print+eval$q,q,a, [download]	[reply] [d/l] [select]
Re^4: Regex lookahead, lookbehind by rahulme81 (Sexton) on Feb 23, 2017 at 11:40 UTC
Hi choroba ... sorry for the confusion with the post. below is what I am doing. `foo_bar_foo10.1.1.1.foo10.1.1.1.txt foo_bar_foo10.1.1.1.foo10.1.1.1.FOO-1.txt foo_bar_foo10.1.1.1.foo10.1.1.1.test_FOO1.txt foo_bar_foo10.1.1.1.foo10.1.1.1.test_FOO_abc_mon.txt` [download] Above are the files which I need to query with regular expression. Out of which I need to pick only file2, file3, file4 and move them to some other directory. file1 should remain untouched. Similarly sometime another set of files is as below `foo_foo_foo10.1.1.1.TEST.failed foo_foo_foo10.1.1.1.test_FOO1.failed` [download] Out of which I need to pick only file2 and move them to some other directory. file1 should remain untouched. As of now I have separte regex to filter and picking the desired files with grep -v. Thanks	[reply] [d/l] [select]
Re^5: Regex lookahead, lookbehind by choroba (Cardinal) on Feb 23, 2017 at 12:18 UTC
Re^3: Regex lookahead, lookbehind by haukex (Archbishop) on Feb 23, 2017 at 10:33 UTC
One last thing I need to extend to below match also When asking for help on regexes, please provide all the possible test cases. Also, you can use a module like Test::More - first, write as many test cases as possible, then work on your regex until all the tests pass, or post the code here if you have trouble. `use warnings; use strict; use Test::More; my $regex = qr/foo/; like "foobar", $regex; unlike "quzbaz", $regex; # more test cases here done_testing;` [download]	[reply] [d/l]