>
>You can probably send several commands at once and use multiple result sets (say, do 10 tables in one command).

For 300+ tables ? Even if all are maximum vfp dbf+dbt capacity, that is below 1.3 TB. There will be a bottleneck somewhere, but not doing it 1 by one will not eliminate that, only add to needed checking - dunno if it might introduce even more overhead at the disk physical level. Race condition would be to harsh, but moving off best sequential read positions, if SQL is not on SSD but traditional RAID....

Run a simple count (*) first to see all tables are responding, save results to compare after moving to vfp and fire off. KISS