Here are some tips if you don't want to use external tools. The key is to be lazy and defer as much work as possible. One simple thing you can do is

    my ($filename, $rest) = split /,/, $_, 2;
    if (index($filename, $to_find) > -1) {
      my ($cms, $path, $size, $day, $time) = split /,/, $rest;
[download]

Going quite a lot further:

  open(F, "+< $infile"); 
  while (sysread F, $_, 32768) { 
    $_ .= <F>;
    next unless /\Q$to_find\E/; # quickskip

    for(grep /^[^,]*?\Q$to_find\E/, split /\n/, $_) {
      ($filename, $cms, $path, $size, $day, $time) = split /,/;
      $href = "file:\\\\netd\\data".$path."\\$filename";
      $href =~ s/\s/%20/g;
      $table .= "<TR><TD><A HREF=\"$href\">$path\\$filename</A><TD>$si
+ze<TD>$day $tim
+e</TR>";
    }
  }
  close(F);
[download]

This greatly reduces the number of IO operations and restricts the heavy splittage to known matches.

I believe the following is an extra win, but I haven't benchmarked it.

  open(F, "+< $infile"); 
  while (sysread F, $_, 32768) { 
    $_ .= <F>;
    next unless /\Q$to_find\E/; # quickskip

    while(/^([^\n,]*?\Q$to_find\E[^\n]*)/gm) {
      ($filename, $cms, $path, $size, $day, $time) = split /,/, $1;
      $href = "file:\\\\netd\\data".$path."\\$filename";
      $href =~ s/\s/%20/g;
      $table .= "<TR><TD><A HREF=\"$href\">$path\\$filename</A><TD>$si
+ze<TD>$day $tim
+e</TR>";
    }
  }
  close(F);
[download]

Beyond these tricks, you quickly get to the point of pretty much handrolling your own database engine..

Makeshifts last the longest.