Re: The BOOLEAN data type
Date: 12 Apr 2003 16:33:46 -0700
Message-ID: <c0d87ec0.0304121533.3cd9aba7_at_posting.google.com>
>> The boolean type has been in the standard for a while, at least
since SQL:1999. But it isn't a _core_ SQL:1999 requirement. <<
To quote from SQL-99 COMPLETELY, REALLY by Gulutzan & Pelzer:
"Warning: by saying that UNKNOWN and NULL are both to be used interchangably to mean the same thing, one is saying that "I don't know" and "I know the data is missing" are the same thing. The drafters of the SQL Standard apparently forgot the distinction, and they have ben justly critisized for this error."
The foundations of SQL asserted that: (1) UNKNOWN and NULL are totally different values in totally domains (2) All datatypes can support NULLs (3) NULLs have predictable values, which include propagation. Look at the basic 3VL:
TRUE OR UNKNOWN = TRUE
TRUE OR NULL = NULL = UNKNOWN -- assuming propagation of null!!
Well, there goes the foundations!!
>> Has that changed in SQL:2003? <<
We can only pray... Received on Sun Apr 13 2003 - 01:33:46 CEST