>>>>>In order to eliminate this issue I am thinking of always specifying/allowing NULL for every column of the empty cursor. Just in case. Will it slow down the processing of this cursor? The cursor usually is not very big, maybe up to 2000 records. What do you think?
>>>
>>>Other than allowing for sloppy programming, there is no conceivable benefit in allowing NULL and there are huge negatives.
>>
>>In some cases null is a must. Consider a lab result, where null means "was not done", and zero means "done, none found".
>>
>
>1. If the lab result is zero ?
Yes, as in "test done, we found not even a trace of what we were looking for".
>2. You mix the process with the result here.
>Correct data require two fields: result and test counter.
Or an individual record for each test, then each counts as one.