I have this regex: ^regex*\.0-9

You do not use code tags, so it's not clear what regex you have. Please see Markup in the Monastery, Writeup Formatting Tips and How do I post a question effectively?.

I going to guess that your regex is  /^regex*\.[0-9]/ You say that your regex matches "regex.1", which I can accept, but why would you think it would match "regex.abc", since "abc" has no relation to  [0-9] (a decimal digit)? Please see perlre and perlretut.

It seems as though the ? doesnt make the \. optional.

I don't see any  ? metacharacter. (Update: And what difference would you expect it would make if the  \. were optional?)

Update: When I first read "It must also match regex, but not regex.abc." in the OP, I thought codz67 wanted the regex also to match  'regex.abc' and was puzzled why it didn't. A more careful reading brings a better understanding. See discussion below. Oh, well...

Something using zero width negative look ahead (?!) should work:

my @t = qw( regex regex.1 regex.12 regex.1a regex.a regex.abc regex.a1
+ oregex.2 );
for (@t) {
    print $_;
    print ' matches!' if /^regex(?!\.[^\d])(\.\d*)?/;
    print "\n";
}
[download]

regex matches!
regex.1 matches!
regex.12 matches!
regex.1a matches!
regex.a
regex.abc
regex.a1
oregex.2
[download]

Checkout the "Look-Around Assertions" section in the perlre documentation. In the example above, the negative look-ahead says fail if the next characters are a period followed by anything that isn't a digit.

Loops: You read the OP a bit more slowly and carefully, and came closer, I think, to answering codz67's question as a result. But the  (\.\d+)? is redundant for the given inputs and outputs if the negative look-ahead is used. (Had it been needed, it would probably have been better as a  (?:...) non-capturing group.)

c:\@Work\Perl>perl -wMstrict -le
"my @t = qw( regex regex.1 regex.12 regex.1a regex.a regex.abc regex.a
+1 oregex.2 );
 for (@t) {
   printf qq{'$_'};
   printf ' matches!' if /^regex(?!\.[^\d])/;
   print '';
   }
"
'regex' matches!
'regex.1' matches!
'regex.12' matches!
'regex.1a' matches!
'regex.a'
'regex.abc'
'regex.a1'
'oregex.2'
[download]

Hi,

Yeah you're right about that. I confess to removing the non-capturing annotation for simplicity sake. Here's a slightly different version where the \d* isn't redundant:

my @t = qw( regex regex. regex.1 regex.12 regex.1a regex.a regex.abc r
+egex.a1 oregex.2 );
say s/^regex(?!\.[^\d])(\.\d*)?/|$&|/r for (@t);
[download]

|regex|
|regex.|
|regex.1|
|regex.12|
|regex.1|a
regex.a
regex.abc
regex.a1
oregex.2
[download]