Re: more on delete from join
From: Mr. Scott <do_not_reply_at_noone.com>
Date: Fri, 28 Aug 2009 20:06:51 -0400
Message-ID: <b5qdnYS2u7KG7AXXnZ2dnUVZ_hCdnZ2d_at_giganews.com>
>>> Mr. Scott wrote:
>>> ...
>>> That's a stretch, it would be more accurate to say that <AND> is to
>>> relational intersection as <OR> is to relational union. If what you
>>> wrote is what most people think it might help explain why so many want
>>> to associate the result of a relational operation with the operations
>>> used for form the result, even though the form of the resulting set of
>>> tuples offers no way to record the operations that were used to produce
>>> it.
>>> I think it is kind of phony for people to appeal at all to algebraic
>>> operations in this way when the strict use of algebra can use <OR> to
>>> produce a relation that is algebraically equal to one produced by <AND>.
>>> That's why I sometimes say a union is always a join, even though it does
>>> seem to wind people up. The expression used to form a view can be said
>>> to persist only if it is recorded as a constraint, which I would say it
>>> should be.
Date: Fri, 28 Aug 2009 20:06:51 -0400
Message-ID: <b5qdnYS2u7KG7AXXnZ2dnUVZ_hCdnZ2d_at_giganews.com>
"paul c" <toledobythesea_at_oohay.ac> wrote in message news:8JUlm.41207$Db2.18172_at_edtnps83...
> Mr. Scott wrote: >> "paul c" <toledobythesea_at_oohay.ac> wrote in message >> news:WFTlm.41198$Db2.23941_at_edtnps83...
>>> Mr. Scott wrote:
>>>> "paul c" <toledobythesea_at_oohay.ac> wrote in message >>>> news:eimlm.40861$Db2.21494_at_edtnps83...
>>> ...
>>>>> I don't know if the plural 'disjunctions' is a typo'. Do you mean >>>>> that the resulting value of R stands for "r1 OR r2, etc." is true? >>>> I don't seem to be getting through. >>>> >>>> AND is to INTERSECTION as OR is to UNION. >>>> ,,,
>>> That's a stretch, it would be more accurate to say that <AND> is to
>>> relational intersection as <OR> is to relational union. If what you
>>> wrote is what most people think it might help explain why so many want
>>> to associate the result of a relational operation with the operations
>>> used for form the result, even though the form of the resulting set of
>>> tuples offers no way to record the operations that were used to produce
>>> it.
>>>
>>> I think it is kind of phony for people to appeal at all to algebraic
>>> operations in this way when the strict use of algebra can use <OR> to
>>> produce a relation that is algebraically equal to one produced by <AND>.
>>> That's why I sometimes say a union is always a join, even though it does
>>> seem to wind people up. The expression used to form a view can be said
>>> to persist only if it is recorded as a constraint, which I would say it
>>> should be.
>> >> I give up. Instead of applying plain old ordinary logic, you appear to >> have strange preconceived notions that are a mystery to me. Without a >> common frame of reference, there's no point in continuing this >> discussion. > > You could look at TTM Appendix A and ask why did D&D distinguish <AND> > from logical AND. Probably they had several reasons, but the most obvious > one would be so that they could avoid circular definitions.
I don't need to look. <AND> is a relational operator; AND is a logical operator. They are related only because the predicate of the result of <AND> is the conjunction of the predicates of its operands. Received on Sat Aug 29 2009 - 02:06:51 CEST