m/filename=(\d.*?)([a-zA-Z].*?)\.j/i)

I think that you probably want \d*? and [a-zA-Z]*?. Note the absence of the '.'.

Ie. You are applying the quantifier (*?) to the '.', not to the character class you want to match.

For starters, Perl and CPAN will help you solve your puzzle:

use YAPE::Regex::Explain;
say YAPE::Regex::Explain->new( qr{/(\d.*?)/i} )->explain;
[download]

----------------------------------------------------------------------
  (                        group and capture to \1:
----------------------------------------------------------------------
    \d                       digits (0-9)
----------------------------------------------------------------------
    .*?                      any character except \n (0 or more times
                             (matching the least amount possible))
----------------------------------------------------------------------
  )                        end of \1
[download]

On other words, at lines 11 and 24, you'll be better off with a simple "\d+" to deal with any number of digits, before moving on to other possibilities.

Semi-OT NOTE: YAPE::Regex's use of "digits" plural ! for a single "\d" seems misleading, to me. Yes, of course the next char, matched by the dot, could be a digit, but that explanation is obscured by the use of the plural in the explanation of \d).

Here's a very loose extrapolation/translation of YAPE::Regex's output in language I think is actually more precise:

----------------------------------------
        \d       one digit <br>
----------------------------------------
        .*?      anything (except a newline), with as few as 
                     zero instances (times) of anything
                     ...and absolutely no more than (in the case of yo
+ur 
                     sample data) 'one additional digit'."
----------------------------------------
[download]

When the best tools you know how to use come up short, you can use a few print statements for debugging many problems -- t'ain't necessarily the best way, but you might find illumination were you to insert just two such:

1. as the first line of the main if -- (print "\t I'm in if \n";)
   ...and
2. another as a new line in the first elsif -- (print "\t I'm in the first elsif \n";).

Obligatory observation: you're not using strict and warnings. You should, at least until you know Perl so well that their assistance is no longer needed to write code that (usually) does what you want and expect... and so well that you know when you can save yourself the line(s) of code.

In case you're wondering at the much-belated and possibly-redundant post, this node has been a work in progress for many hours -- interspersed with Fire calls (2 each), SWMBO's decision (1 each) that this was the day to select and cut an Xmas tree, and loading/delivering a cord of wood (1 each, $350/cord) with a friend... Probably not much point in saying more; maybe not even in this, at this late date.

Unfortunately \d is not always [0-9]. Under unicode it includes rather more codepoints than 10. If you are writing validation logic it is better to not use \d and use [0-9] instead.

Good point. ++

@Files =( "filename=merilyn14.jpg.jpeg", "filename=...)

$file = "filename=merilyn14.jpg.jpeg"
[download]

mmmh... That seems a strange way to me to call a file, a source of potential troubles with your script in the future, "filename=merilyn14.jpg.jpeg" is not a file, nor even a filename, I will fix this if I were you.

more questions

Maybe you will prefer this, probably less ambiguous and more readable, form...

And

foreach my $file (@Files)
      {
          if ($file =~ m/filename=(\d.*?)([a-zA-Z].*?)\.j/i)
[download]

... same as:

foreach my $file (@Files)
      {
          if (/^filename=(\d.*?)([a-zA-Z].*?)\.j/)
[download]

Don't need to use /i in the match if you are using also a-zA-Z

++ for the rest, but "wrong" to your final comment:

The /i has no impact with the character class ... but may very well have to do with the "j". If OP was concerned about the various combinations of .jpg, .JPG, Jepg, JPEG and so on, then the /i is, in fact, very much needed.

Updated to clarify the last sentence above

I didn't see the extension. Welldone, smart note...

But, although you are pointing in the right way, in this particular case could be a false problem...

Remember, this is what we are really feeding to the script.

@Files =(
"filename=merilyn14.jpg.jpeg",
"filename=003_Merilyn23.jpg.jpeg",
"filename=890FoxyM.jpg.jpeg",
"filename=006.jpg.jpeg"
);
[download]

At least in some operative systems "filename=merilyn14.jpg.JPEG" will never be evaluated or modified by the regexp, because is not in the foreach loop, so be aware of this point. We can safely avoid /i here if we are in Linux, cause its our duty to check before what we put in @Files, and even if it was one or several JPG we should change only the regexp for those matches. There is not point into having a regexp searching something that we know will never match.

You say "/i has no impact with the character class". Which in this case is true as the class is [A-Za-z] but is not generally true:

perl -Mre=debug -e'/[a-z]/i'
Compiling REx "[a-z]"
Final program:
   1: ANYOF[A-Za-z][] (12)
  12: END (0)
stclass ANYOF[A-Za-z][] minlen 1 
Freeing REx: "[a-z]"
[download]

use strict;
use warnings;

my @files = qw|
filename=merilyn14.jpg.jpeg
filename=003_Merilyn23.jpg.jpeg
filename=890FoxyM.jpg.jpeg
filename=006.jpg.jpeg
|;

my ($file, $folder);
for (@files) {
    next if !s/^filename=//;
    s/(\.jpe?g)+$//;
    $file = "$_.jpg";
    s/^\d+_?//;
    s/_?\d+$//;
    $folder = $_ || 'FileNum';
    print "$file -> $folder\n";
}

__DATA__
OUTPUT:

merilyn14.jpg -> merilyn
003_Merilyn23.jpg -> Merilyn
890FoxyM.jpg -> FoxyM
006.jpg -> FileNum
[download]