Re: Regex: Identifying comments

[Athanasius]

in reply to Regex: Identifying comments

Here is another approach:

#! perl
use strict;
use warnings;

my @selects = ("select 'text' from foo        --This is a comment",
               "select '--Not a valid comment' from foo        --But t
+his is",
               "select '--This is not a valid comment' from foo",
               "select '--Not this' + '--either' from foo");

for my $select (@selects)
{
    my ($comment) = $select =~ / ( -- [^']+ ) $ /x;
    print $comment, "\n" if $comment;
}

Output:

--This is a comment
--But this is

Update: The above code does not handle single quotes in comments. The following is one way to do this (indebted to choroba’s solution, above):

#! perl
use strict;
use warnings;

my @selects = ("select 'text' from foo        --This is a comment",
               "select '--Not a valid comment' from foo        --But t
+his is",
               "select '--This is not a valid comment' from foo",
               "select '--Not this' + '--either' from foo",
               "select stuff from that -- a 'useful comment' goes here
+?",
               "select 'qaws' + make from \"a\" -- comment with 'a' qu
+ote",
               "select 'a' from 'b' with 'c' -- comment with 'a --' co
+mment");

for my $select (@selects)
{
    my $stripped = $select =~ s/ ( ' .*? ' ) /'_' x length $1/egrx;

    if ($stripped =~ /--/)
    {
        my $i = length($select) - (index reverse($stripped), '--') - 2
+;
        print substr($select, $i), "\n";
    }
}

Output:

--This is a comment
--But this is
-- a 'useful comment' goes here?
-- comment with 'a' quote
-- comment with 'a --' comment

Hope that helps,

[RichardK]

I like the look of this approach, but is this a valid comment ?

 select stuff from that -- a 'useful comment' goes here?

[zuma53]

Yes, as the quote comes after a free-and-clear '--'

[choroba]

See the update to my reply which should handle this. It cannot handle multiline string in single quotes, though.