Re: Interesting question and debate on ddl tables with date fieldsthat will not always have a value

Walden H. Leverich wrote:

Also keep in mind, there is a practical issue here too. Many systems
(including query and ETL tools) understand that 'null' means something
special, specifically that there is no value for the column. If you
decide to load up the date column with 0001-01-01 then you'll have to
special case a lot of code that would otherwise just understand that
there was no value for that column.

Your post is theoretically correct, Walden, but think about it practically: if you are indeed using a status column, try to name a situation where you would ever use the "nullness" of the field. I've stretched my brain, and I can't think of one. Whether the terminate date has null or 0001-01-01, I won't be testing it either way, because I will have already tested the status field.

On the other hand, from a purely practical standpoint, if you use null, you will now have to specify *NULLIND for every place where you pass that field to another procedure. Is it worth the tradeoff? I don't know.

Anyway, thanks for an enlightening discussion. I'm *not* an avidly anti-null guy. I just have found few uses for it. I enjoyed the civil discussion; they're increasingly rare these days both on the list and in the "real world".

Joe