OK. I knew that was the right thing to do. I was just hoping I could somehow cheat. It means a little re-writting, but when finished, it will be the RIGHT way.

Thanks for the constructive criticism.


>Matt,
>
>I think it's time for you to break down and add a sequence number to the child table. The order of records returned by SELECT should be considered random unless they're ordered with ORDER BY. As you've seen, different SELECT statements will order them differently. Plus, if you ever up-size to a client/server solution will break any code that relies on physical record order because most relational databases don't have any concept of physical record order.
>
>Do your self a favor and add sequence numbers to the child table and order it with ORDER BY.
>
>Just my $0.02.
>
>Josh
>
>>I have a SQL select with a simple join.
>>
>>When completed, the records are in the reverse order for matching children keys. In other words, I have an 'order by' statement that orders by the child-key, and that part works fine, but within each group of child-key records, they are now in the reverse order from what they were originally entered into the table.
>>
>>This does NOT happen if I eliminate the join step.