Re: more on delete from join
Date: Wed, 26 Aug 2009 20:03:04 GMT
Tegiri Nenashi wrote:
> On Aug 26, 10:40 am, paul c <toledobythe..._at_oohay.ac> wrote:
>> Insert to a base relvar represents a conclusion: R =: R UNION A conludes >> R AND A.
> What is "R AND A"? Some proposition (aka 0-ary predicate)? This can't
> be because join of R and A certainly is not 0-ary.
I meant "AND" as the logical connective, not shorthand for relational "<AND>". "R AND A" stands for a logical conjunction of propositions, each proposition having been concluded to be true. Relational algebra has no form to record, say, "proposition 1 (R) is true OR proposition 2 (A) is true" in one relation, other than to interpret each of those two propositions as a relation with a single tuple, eg., R <OR> A, ie. two relations. in a single relation, only "proposition 1 is true AND proposition 2 is true" can be recorded. So either proposition might be false but when a relation represents their predicate(s), both are interpreted to be true (because they can be 'queried' individually giving a query result for either and both of 'true'). That's why I say we can conclude R AND A from R := R UNION A (or from R <OR> A).
Granted, this might look like mixing up logical connectives with relational operators, nevertheless the conventional interpretation of each results in a conclusion about both, as disturbing as it might seem.
This seems to me to be an effect of combining two factors: 1) language convention, namely the devices of 'insert' and 'delete' and 2) the conventional relation form. A different relation form might allow insert to be defined in terms of <AND> rather than <OR>, but that's another question. Received on Wed Aug 26 2009 - 22:03:04 CEST