Interesting observation. I always thought that all VFP procedural commands respect current index order. I just checked REPLACE FOR and it doesn't respect index order either if FOR condition is not optimizable or affects whole table
>This may be a surprise.
>
>xbase commands should respect an active index,
>but it is not always true.
>
>With REPLACE ALL the write process run with the RECNO() order.
>
<snip>
--sb--