First of all, you can often get the database itself not to accept any duplicate numbers, if you define that table with a unique index over the number. That will work much better than doing it on your side, particularly when the table gets big.

Now to your code.

• When you're looking for a number like that, you don't need to use $&, simply enclose what you're looking for in ().
• You don't need $', because Perl will keep track of where it was in the string.
• You don't need to scan continously over an array to see if a number is there, that's what hashes are for.
• You don't need the GOTO, because Perl provides the next keyword.

I left your variable names as they were, so that you will find it easier to compare your code to mine. But in general, all caps for a variable name is rare in Perl. Oh, and contends should probably be contents...

while ($CONTENDS=~/(\d{10})/g) {
    $match=$1;
    $IDNUMBER{$match}=1 unless $IDNUMBER{$match}; 
}

foreach $match (keys %IDNUMBER) {
  $sqlinsert="INSERT INTO $tab[$loop] VALUES('$match')";
                 $rc=$db->Sql($sqlinsert);
                die qq(SQL fail "$rc":), $db->Error(),qq(n) if $rc;
}
[download]