We had about 36 queries, many update, delete and append queries. When it seemed to be a particular query we would test by having the process skip that query, it then just became corrupted on the next query.
SO, we divided the queries into three sets, A, B and C.
We tested and found that regardless of the order executed, either A-B-C or A-C-B or B-A-C or B-C-A or C-A-B or C-B-A
the corruption would always take place in the second set. It seemed that it was not the specific query that was the issue, but some issue which involved an accumulation affect, (maybe in some communication/network buffer?)
Now this corruption would only work when there were at least two in the database thru the network. (Though there were certain conditions when the corruption would not take place depending primarily on whether the one running the queries was the first one in the database.)
In any case, instead of using the workarounds, we worked it out so those processes would only be initiated with one person logged in, since it was just needed once in the morning. This took care of 95% of the corruption issue; there would still be an issue once or twice a month.
NOW, for the past 5 months we have had no corruptions. So, to test it, we allowed for the series of queries to be run with others logged in as well, and there is no longer a corruption issue. (I assume there has been some Microsoft update, may be relating to networking, that inadvertantly has also corrected the corruption issue.)