I can think of a few ways of doing this -

Method 1 (row by row comparison, the code is not tested, just a general idea of how the code would look like)

#!/usr/bin/perl -w
use strict;
use DBI;
use DBD::Oracle;

my $db1 = DBI->connect("dbi:Oracle:server=XXXX;database=XXXX", "USER",
+"PASSWD")
        or die "Failed to connect to the database!";

my $db2 = DBI->connect("dbi:Oracle:server=XXXX;database=XXXX", "USER",
+"PASSWD")
        or die "Failed to connect to the database!";
        
my $table = "DUMMY";        # name of the table
my $unique_id = "ID";       # unique id to identify the row

my $sth1 = $db1->prepare( "select * from $table" );
my $sth2 = $db2->prepare( "select * from $table where $unique_id=?" );

$sth1->execute();
while (my $row = $sth1->fetchrow_hashref())
{
    my $found = 0;

    $sth2->execute( $row->{$unique_id} );
    while (my $row2 = $sth2->fetchrow_hashref()) {
        # compare the row
        foreach (keys %$row) {
            if ($row{$_} ne $row2{$_}) {
               print "DIFF: (ID) $_\n";
               break;
            }
        }
    }
}

$sth->finish;
$dbh->disconnect;
[download]

Method 2 - DBA way

step 1 - copy both tables into a dummy database, name the tables dummy1 and dummy2

step 2 - copy dummy1 to new table dummy3

step 3 - SQL: delete from dummy1 where ID is found in dummy2

step 4 - delete from dummy3 where ID is found in dummy1

step 5 - delete from dummy2 where ID is found in dummy 3

So in the end, dummy3 holds records in both the original dummy1 and dummy2 tables, and dummy1, dummy2 hold records that do not reconcile.

Method 3 - pg way

Have to say that pg's method is much clever.