Path: newssvr20.news.prodigy.com!newsmst01.news.prodigy.com!prodigy.com!pd7cy2so!shaw.ca!elnk-pas-nf1!newsfeed.earthlink.net!stamper.news.pas.earthlink.net!newsread2.news.pas.earthlink.net.POSTED!90a49e3f!not-for-mail From: Jonathan Leffler Organization: Randomly Disorganized User-Agent: Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.5) Gecko/20031007 X-Accept-Language: en-us, en MIME-Version: 1.0 Newsgroups: comp.databases.theory Subject: Re: Database-valued attributes? References: <_C_rb.122225$9E1.607217@attbi_s52> In-Reply-To: <_C_rb.122225$9E1.607217@attbi_s52> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Lines: 49 Message-ID: Date: Tue, 11 Nov 2003 06:36:24 GMT NNTP-Posting-Host: 66.47.74.12 X-Complaints-To: abuse@earthlink.net X-Trace: newsread2.news.pas.earthlink.net 1068532584 66.47.74.12 (Mon, 10 Nov 2003 22:36:24 PST) NNTP-Posting-Date: Mon, 10 Nov 2003 22:36:24 PST Xref: newssvr20.news.prodigy.com comp.databases.theory:22602 Marshall Spight wrote: > 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? Hmmm...would the elements of the set of tables all need the same type? In the RMD (relational model of data), the values in a column are homogeneous in type, usually. For example, in a table with one or more RVAs, the RVAs for a given attribute have the same structure/type in each row. What is a database? That isn't intended to be wholly facetious. If a database is the entire universe of discourse, then you can't have a database embedded in anything else. If it is an organized collection of relvars and constraints, then it is possible to do more with them, including potentially embedding database values in some other relational structure. > 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. > > If one wanted to have an attribute that belonged to the tree domain, > one would need to support an attribute with such a pair of relation > values. So, this would be a pair of relations with the same type in each row? Or is it one row per tree, with potentially different relation types for each row? I worry about the latter; the former is probably manageable in some way - I'm not sure we have the jargon yet, but there is a structure there... > Does this raise any problems? I don't see any immediate issues, but > it's not an idea I've ever heard mentioned anywhere else. > > Comments welcome, including alternative ideas. I've probably not spent enough time digesting the concept, so accept the questions as superficial and possibly misguided - but they may perhaps help you understand what you're dealing with. -- Jonathan Leffler #include Email: jleffler@earthlink.net, jleffler@us.ibm.com Guardian of DBD::Informix v2003.04 -- http://dbi.perl.org/