Re: WWW/Internet 2009: 2nd CFP until 21 September x

From: rpost <rpost_at_pcwin518.campus.tue.nl>
Date: Sun, 16 Aug 2009 20:12:10 +0000 (UTC)
Message-ID: <h69p6q$2no2$3_at_mud.stack.nl>



Bob Badour wrote:

>Walter wants a clue. If every table with null represented a materialized
>outer join, as he claims he thinks it does, then presumably something is
>being joined. Unnamed perhaps. Implicit for sure. But some ephemeral
>tables or relations are joined (outer joined) to materialize the table.

Given a relation with a key K1, ... Kn and remaining attributes L1, ... Lm, that relation is always the join of the m projections on K1, ..., Kn, L1, ..., K1, ..., Kn, Lm. What Walter means is that we can describe 'missing information' NULLs (not all NULLs, mind you) as the result of allowing such projections to have fewer rows than the full relation.

This describes tables with NULLs as shorthands for sets of tables without any NULLs at all, so it introduces them in a way that fits perfectly into the relational model.

If you actually had a clue, rather than just pretending, you wouldn't call these projections 'unnamed', 'implicit' or 'ephemeral'.

-- 
Reinier
Received on Sun Aug 16 2009 - 15:12:10 CDT

Original text of this message