After talking to my boss she thinks the two customers that this has happened to have fairly slow networks (well below the spec our software requires). Is it a possibility that in the small fraction of time that one request to overwrite the view has deleted the records and the next request is checking the DBC that two users are actually writing the same view to the DBC but with different definitions and different unique ids for the parentid??? It appears that the two views were consecutive in the uniqueids ..one has parentid of 835 and the other has parentid of 1036... all the ids between 835 and 1035 are used by the first view for fields. It this is a possibility than I guess I'll be writing a locking routine to force the second user to wait.

Thanks again