No NULLS ins SAS!
Although some defend that NULLS and MISSING are conceptually the same, they do represent two different things.
From my understanding:
NULL value should mean, NO VALUE FOR
MISSING value should mean, NO KNOWN VALUE FOR
First assumes that some particular information has no value for some attribute, the second assumes that there is no information. Clearly two different things for data analysis and statistics.
Another difference is that a MISSING value represents actually a numeric value, it evaluates to - infinite. NULLS are just NULLS and can only be compared as a match to another NULL value or not.
Anyway, SAS will translate MISSINGS to NULLS when loading data to a RDBMS.
Being a datetime a numeric value, just use the missing notation (represented by . ) like this:
[pre]
select . as DT format=datetime30.6 ...
[/pre]
Cheers from Portugal.
Daniel Santos @
www.cgd.pt