Because of LOB processing, PowerExchange has reduced the target array size to a number smaller than the source array size. To compensate, PowerExchange has reduced the number of rows in each source request.
Processing continues with fewer rows in each request.