Corrupted SE version 5.0 table

I'm using SE version 5 and have a 350,000 row table. I try to use esql/c with a delete from table to remove all the rows. It took about 12 hours of processor time on a multi-processor NCR 3555 to delete all the rows. It took about 1 hour to insert the rows using individual inserts from esql/c. I can drop the full table, create it, load it again and my delete goes OK. I think the problem the first time was that my machine got re-booted in the midst of the either the delete operation on the insert operation. When I reloaded the table, the idx file was about 30 % smaller. I have a feeling that the reboot did a number on the .idx file causing some kind of a corruption that was not found by a bcheck. Does this sound familiar to anyone. I'm guessing here but can there be corruptions that bcheck can't find. Thanks in advance John Estal