Path: newssvr20.news.prodigy.com!newsmst01.news.prodigy.com!prodigy.com!in.100proofnews.com!in.100proofnews.com!newsfeed.cs.utexas.edu!geraldo.cc.utexas.edu!an02.austin.ibm.com!ausnews.austin.ibm.com!newsfeed.btv.ibm.com!gazette.almaden.ibm.com!not-for-mail From: "Paul Vernon" Newsgroups: comp.databases.theory Subject: Re: Database-valued attributes? Date: Tue, 11 Nov 2003 12:06:50 -0000 Organization: IBM Almaden Research Center Lines: 35 Message-ID: References: <_C_rb.122225$9E1.607217@attbi_s52> NNTP-Posting-Host: pwcukvsg3.emea.ibm.com X-Trace: gazette.almaden.ibm.com 1068552608 41652 9.133.41.65 (11 Nov 2003 12:10:08 GMT) X-Complaints-To: postmaster@almaden.ibm.com NNTP-Posting-Date: Tue, 11 Nov 2003 12:10:08 +0000 (UTC) X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2800.1158 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Xref: newssvr20.news.prodigy.com comp.databases.theory:22612 "Marshall Spight" wrote in message news:_C_rb.122225$9E1.607217@attbi_s52... > There's been the occasional incidental discussion of relation-valued > attributes on the list lately. This has got me thinking: what about > database-valued attributes? That is to say, what about an attribute > that consists of a set of relation values? > > The reason I ask is because I have been thinking about expressing > trees as values. The canonical logical representation of a tree from > "Practical Issues in Database Management" is a pair of relations: > one for nodes, one for edges. Bingo. I think you will find that canonical logical (i.e. relational) representations for the vast majory of interesting data structures (i.e. not just trees) consist, in each case, of greater than one relation. So yes, we would indeed want database valued attributes. Just need to agree on what a database value actually is. If it is a 'set of relations' what happens if two relations contain the same set of tuples (of the same type), but have different 'relation names'? Are they different relations? If so does that mean that if *relation variables* have names, then *relation values* must also have names? I.e. I suggest that either neither relation values and relation variables have names, or that both do. I suggest we cannot have either-or. :-) Regards Paul Vernon Business Intelligence, IBM Global Services