Excluding pattern in regular expression search and replace - regex

Question

I have this string
asp.net somedomain.com
I need to strip out the domain dot extension part only except in certain cases. So I want this:
asp.net somedomain
Any time there is vb.net, asp.net etc.. I do not want to strip out the extension.
I tried this in perl with no effect.
$company =~ s/(?=\w+)(?!=asp|vb|c#)\.[a-zA-Z]{2,6}\b/\1/g;
My logic is stuff before the dot must be one or more alpha and not asp or vb or c#.

Answer 1

You can use a Negative LookBehind. You were almost there, but using LookAheads.
RegExp: (?<!asp|vb|c\#)\.[a-zA-Z]{2,6}\b
Replace with nothing
Explained demo here: http://regex101.com/r/tG5rO1
To workaround the variable length RegEx error use this: (?<!asp)(?<!vb|c\#)\.[a-z]{2,6}\b
Edit: separate LookBehind group for different length excluded word
This will only find TLD's that don't match your criteria.
Update:
To take care of special cases: don't match words ending in excluded word and match any combination of excluded word (eg. vB VB vb Vb)
RegExp: \b(?<!\b[aA][sS][pP])(?<!\b[vV][bB]|\b[cC]\#)\.[a-zA-Z]{2,6}\b
Explained demo: http://regex101.com/r/bR3kJ8
Or: \b(?<!\basp)(?<!\bvb|\bc\#)\.[a-z]{2,6}\b
When used with case insensitive RegEx modifier i
Update #2
Safer as it cares only about .net TLD and excluded words for it:
/(^|\s)(?!(?:visual)?(?:basic|studio|asp|v[bs]|c\#)\.net)(\w+)(?:\.com?\.[a-z]{2}|\.[a-z]{2,6})\b/\1\2/gi
Needs replacement as opposed with previous variants.
Explained demo: http://regex101.com/r/kL5mQ5

Answer 2

Just match the last one:
my $s = q{asp.net somedomain.com};
my ($company) = ($s =~ / ([A-Za-z]{2,}) [.] (?:[A-Za-z]{2,}) \z /x);
print $company, "\n";
Or, split on space and dot:
my $s = q{asp.net somedomain.com};
my ($company) = split /[.]/, (split ' ', $s)[-1];
print $company, "\n";
How much work you want to put into the pattern depends on how much variation there is in your input. The examples above are based on the sample input your provided.