Hello,

I found this article interesting. However, granted I'm not a full-time DB designer, I am slightly confused by the statement:

For existing databases (in cases where I can't design the table from the ground up) I prefer not to use some arbitrary value to denote "no value" and do use NULL values in this case. I find using LEFT JOINs or searches WHERE x IS NULL easy enough in these cases...maybe I'm missing something?

However, I do agree that it's always best to extract the columns out into their own table - removes the problem completely.

Is there an article/source that explains why it is better to use a predefined value instead of NULL?

Thanks, George

.NET Programmer Analyst